1 1 HEALTHY INCENTIVES PILOT (HIP) SYMPOSIUM 2 U.S. DEPARTMENT OF AGRICULTURE, SNAP 3 * * * 4 USDA-FNS-FSP 5 SECOND FLOOR 6 3101 PART CENTER DRIVE 7 ALEXANDRIA, VIRGINIA 8 THURSDAY, OCTOBER 16, 2008 9 9:05 a.m. 10 11 The above-entitled matter convened on 12 Thursday, October 16, 2008 at 9:05 a.m., Steven 13 Carlson, Director, Office of Research and Analysis, 14 FNS, presiding. 15 16 17 18 19 20 21 22 2 1 P R O C E E D I N G S 2 (9:05 a.m.) 3 MS. SHAHIN: Good morning. Could I 4 encourage everybody to take a seat, please, and we'll 5 get started. Let me just mention that we do have an 6 interpreter with us today, so if anyone is in need of 7 an interpreter, please -- probably, it would be 8 easier for you to sit on this side of the audience. 9 Is there anyone here who needs an 10 interpreter? 11 (No response.) 12 MS. SHAHIN: Thank you very much. 13 Welcome. This is such a great crowd and we are so 14 excited to have you all here. It's wonderful to see 15 new faces, to me, but also some very familiar faces. 16 I've had the chance to talk to Jan Walters 17 from Iowa. I think so much of the work she has done 18 in farmers markets, with EBT, and, of course, 19 Lorelei, Forester. Where are you girl? There she 20 is. 21 Dottie Rose is around here, and just a 22 great crowd of people, so to those of you whom I 3 1 know, thank you for being here, and those of you I 2 have not met, I'll look forward to meeting you today 3 and in the future. 4 This is really a wonderful opportunity 5 that we have to hear from the experts. You guys in 6 the room, you're the experts whether we're talking 7 about nutrition, the retail community, folks who 8 understand the technology that has brought us such a 9 long way. 10 And it gives us an opportunity to 11 something very new, and look at ways to really 12 affect the eating habits of people who are eligible 13 for SNAP. 14 I think everybody in the room probably 15 knows, but just in case you don't, we have a new 16 name for the Food Stamp Program, which is the 17 Supplemental Nutrition Assistance Program, and we 18 have little things going on all around the Agency, 19 so that if you say -- I hope this doesn't offend 20 anybody -- if you say the F-word, that means Food 21 Stamps. 22 (Laughter.) 4 1 MS. SHAHIN: And we have to pay, because 2 we're SNAP now. We're really proud of that new 3 name. 4 The thing that you all in this room are 5 interested in, is that this is a program that 6 provides nutrition assistance to people in need. It 7 puts healthy food within reach. 8 What else can we do about that now that 9 we have an opportunity? Our Deputy Under Secretary 10 Kate Houston, is going to talk more about that, and 11 will be introducing and kicking off this meeting. 12 Then we're going to look forward to 13 hearing from all of you. Now, what I will do, is 14 take the opportunity to introduce our Deputy Under 15 Secretary, Kate Houston. She has been a champion. 16 She is a nutritionist herself. In fact, 17 she made it real clear to us, sometimes, that, you 18 know, we're talking about "healthful," not 19 "healthy." She knows what she's talking about. 20 She understands the potential for this 21 program, and she has been a champion for exactly 22 that, and she was one of the people that was right 5 1 there, making sure that this particular provision, 2 which she will talk about a little bit more -- it 3 was in the Farm Bill and made it into law, and I am 4 proud to introduce our Deputy Under Secretary, Kate 5 Houston. 6 (Applause.) 7 UNDER SECRETARY HOUSTON: Thank you all 8 so much. I'm really excited and privileged to be 9 able to participate in this meeting today. 10 I think it's going to be a really 11 exciting time for the Agency as we move forward with 12 what we are now, I guess, calling the HIP Project, 13 the Health Incentives Pilot Project. 14 As Jessica just mentioned, it was 15 authorized in the 2008 Farm Bill, and there is 16 funding that will allow us to test innovative 17 approaches to reducing obesity among participants, 18 specifically by using incentives at the point of 19 sale. 20 There are many benefits of eating a diet 21 consistent with dietary guidelines and the Pyramid. 22 Unfortunately, the diet of most Americans, including 6 1 a lot of us in this room, fall short. 2 I don't need to preach to the choir about 3 the nation's most pressing challenges in poor diet 4 and inactivity, but, low-income individuals, as we 5 also know, are particularly at risk. 6 We are failing to consume adequate levels 7 of certain foods -- fruits and vegetables, for 8 instance -- which we know contribute to good health. 9 But it's not for lack of trying, and it's 10 definitely not for lack of eating. Over 17 percent 11 of children in this country, are overweight, and over 12 66 percent of adults are overweight or obese. 13 The Federal Nutrition Assistance Program 14 that we administer here at FNS, are really powerful 15 tools to address the problem. Nutrition assistance 16 programs service at least one in five Americans every 17 year. 18 We really have an opportunity to help 19 these programs to get in touch with people who need 20 food benefits, but also nutrition education. 21 SNAP alone reaches 29 people each month, 22 and the numbers are growing. We are here today so 7 1 that we can discuss the unique and really exciting 2 opportunity to influence these families and 3 encourage them to purchase and consume more 4 healthful foods. 5 For several years now, we've been 6 discussing the need to support new and innovative 7 ideas that empower low-income Americans, to consume 8 diets that include optimal levels of fruits and 9 vegetables, whole grains, and other foods. 10 The concept of improving the diet of 11 individuals receiving food assistance, alone, is not 12 really new. In fact, each of our 15 nutrition 13 assistance programs, integrate some form of 14 nutrition education into the benefits we make 15 available to participants. 16 Combined, these programs are the largest 17 single funder of nutrition education in the nation, 18 somewhere around $700 million every year. That's 19 provided mostly in the form of grants to state 20 agencies, and, working together with states and many 21 of our partners, nutrition education is now a 22 prominent component of SNAP. 8 1 It can be a powerful tool for empowering 2 the consumer's in food choices, but, given the 3 extent of the problem, education, alone, may not 4 alter the course of this critical public health 5 problem. 6 That kind of brings us to where we are 7 here today. Over time we've had numerous 8 conversations in various forms, regarding how best 9 we could influence the diet of Americans. 10 There are a range of views that exist on 11 this matter, and I hope we'll talk about some of 12 them today. 13 We have long asserted that incentives 14 encouraging the purchase of healthful food, is 15 likely to be more effective in dietary improvement, 16 than punitive approaches. 17 The Administration's Farm Bill proposal, 18 which Jessica alluded to, had requested funds that 19 would allow FNS test incentives and other strategies 20 that may increase the purchase of nutritious food by 21 SNAP participants, through a pilot effort. 22 While the funding available, is far less 9 1 than the one we originally proposed, we are very 2 grateful for the support of Congress, in allowing us 3 to take this important first step, which is also, I 4 think, a historical first step in exploring 5 incentives to improve the diet of SNAP participants. 6 Through this project, FNS will test the 7 use of point-of-sale incentives to determine if the 8 total amount of fruits and vegetables and other 9 healthful foods purchased by SNAP participants, 10 increases. 11 The ultimate goal is to be able to 12 improve the diet of SNAP participants, sort of our 13 bottom line, and whether you're on the panel or just 14 an audience participant, we are really looking 15 forward to hearing from you, your ideas and 16 experience and how they might contribute to our 17 total understanding of the problem. 18 Our partnerships are so critical to the 19 success of our programs, and we're grateful for the 20 information and assistance that you can provide to 21 us, as we devise planning, implementation, and 22 evaluation of the HIP project. 10 1 I just want to say again, how grateful we 2 really are to all of you. This will be a long 3 process and this is the first in what will probably 4 be several meetings and hundreds of discussions 5 about how we can make HIP work. 6 We are truly committed to making sure 7 that we do this right. We have what could be really 8 our only opportunity here, to make sure that we do 9 get it right. 10 Ultimately, we want to make sure that by 11 the end of this Project, we do have some information 12 that can help inform policymaking down the road. 13 With that, I just want to thank you 14 again, all of you, for your time and energy and for 15 being here, and your willingness to provide 16 expertise to us, and for being part of this 17 important dialogue. 18 I'm going to try to stay for as much of 19 the morning as I can. I'm very interested in 20 hearing from our distinguished panels, and I'll be 21 around for conversation, hopefully at the break. 22 Thanks again. 11 1 (Applause.) 2 MR. CARLSON: Thank you, Kate. My name 3 is Steven Carlson, Director of the Office of 4 Research and Analysis here, and I am your host for 5 the rest of the day. 6 Get used to this face. It won't get any 7 better as the day progresses. 8 (Laughter.) 9 MR. CARLSON: I want to start with a few 10 housekeeping details, before we begin with the first 11 major presentation. 12 First of all, you should all be aware 13 that we are recording the proceedings this morning 14 and this afternoon, so we'll have a complete 15 transcript to inform us all of the discussion that 16 takes place. 17 That does mean that all of us, whether at 18 the table or in the audience, really need to be as 19 conscientious as we possibly can, to use the 20 microphones and speak clearly and project loudly 21 enough that the transcriber can pick up on what we 22 have to say. 12 1 Perhaps most important of all, because 2 it's a long day, the restrooms are out in the 3 elevator lobby; ladies to the left, men to the 4 right. 5 It does require a pass. We have a 6 limited number of passes. For those of you who 7 behave -- 8 (Laughter.) 9 MR. CARLSON: -- they will be found out 10 at the registration table. We only ask that when 11 you come back, you turn it back in, so the next 12 desperate person can take advantage of it. 13 The food arrangements and refreshment 14 arrangements in this building and in this area, are, 15 admittedly, limited; you need to know that. But 16 there is a small cafeteria and convenience store 17 downstairs in this building. 18 There are also a few vending machines in 19 the elevator lobby on the fourth floor, that are 20 accessible. They dispense softdrinks and a few 21 snack items and so forth. We will be taking a 22 break for lunch. We will break at 11:30, because 13 1 choices are limited. 2 You may want to go to the cafeteria. 3 It's not the greatest place, but is a convenient 4 place. There are a few restaurants up and down King 5 Street, Route 7, right out front of the building, if 6 you brought a car. 7 There really isn't much within walking 8 distance. 9 We will ask that you return promptly. 10 We'll start the afternoon panel promptly at 12:45, 11 so we can stay on schedule and work our way through 12 the agenda. 13 At this point, I want to introduce to 14 you, Kathy Roark. Please come up. Kathy is going 15 to speak to us about a relatively recent report that 16 the Government Accountability Office completed, that 17 examined what was known and what was to be done on 18 some of the issues that surround the whole concept of 19 providing an incentive to purchase healthful foods. 20 Kathy? 21 MS. ROARK: Thank you. I'm not used to 22 holding a mike. I will get used to it. Wave 14 1 wildly, if you can't hear me. 2 Good morning. Thank you for inviting me 3 for GAO here today. I'm going to provide a brief 4 overview of the GAO report that was published in 5 July of this year, looking at options for 6 incorporating financial incentives for purchasing 7 what we're calling targeted foods. 8 Whatever food are targeted for promotion, 9 such as fruit and vegetables, and the options of 10 providing incentives to purchase those foods, is the 11 SNAP program. I brought my wallet, if I say Food 12 Stamp Program, which I probably will, but I'll just 13 let you know that I can pay. 14 (Laughter.) 15 MS. ROARK: There's a handout on the 16 table over there and copies of that report. I'm 17 going to cover the main issues that I consider most 18 relevant to your expert panel discussion today. 19 First, the requestor of this work, was 20 Senator Tom Harkin, Chairman of the Senate 21 Agriculture Committee. 22 The key research objectives, were to 15 1 understand what we know about the effectiveness of 2 financial incentives and other nutrition initiatives 3 approaches to nutrition education. 4 The second objective was to understand 5 the options for delivering financial incentives, and 6 then the third was to understand what factors should 7 be considered in designing a financial incentive 8 program. 9 The methods we used to assess these 10 objectives, included interviewing a lot of different 11 groups, talking to people, large and small retailers 12 and retail associations, FNS officials, states, 13 including some states with WIC EBT experience, and 14 the main EBT contractors. 15 We also had an expert panel, I believe, 16 about a year ago, and our expert panel was focusing 17 on technical and operational issues. Some of you in 18 the room, are experts. 19 So, the three major areas I want to 20 discuss, are: The definition of healthy foods; 21 options for delivering financial incentives; and 22 then monitoring and evaluation. 16 1 In the handout, there are some page 2 numbers that refer to what more detail you can find 3 in the report. 4 The first issue is the definition of 5 healthy foods. No matter who we are speaking with 6 about this idea, promoting healthy eating by 7 providing financial incentives to SNAP participants, 8 the first question was, well, what's a healthy food? 9 And the second question was, who decides what's a 10 healthy food? It sounds like fairly simple questions 11 with simple answers. It turns out it's not so 12 simple. 13 Many of you are nutritionists, I think, so 14 I don't need to tell you that many food items contain 15 many different nutrients that affect health, and it's 16 not always easy to identify if these items are or are 17 not healthy. The stakeholders we interviewed wanted 18 to have it kept as simple as possible, even though 19 it's not so simple. They didn't want their cashiers 20 or grocery store clerks to become the food police, 21 because of confusion over what's eligible for a 22 discount and what's not. In general, people wanted 17 1 a general category of foods, such as fruits and 2 vegetables, to be identified, rather than a single 3 item, based on nutrient content. 4 The availability of food was also a 5 consideration. Many employees spoke to us and said, 6 well, fruits and vegetables are chosen, but all 7 fruits and vegetables should be promoted, and fresh 8 and frozen, not just fresh. Different areas of the 9 country have different levels of access to those 10 throughout the year. 11 Moving to options, we found the various 12 options in three different areas. The implications 13 of each option varied in terms of implementation, 14 implications and costs. We divided them up into 15 paper, separate cards, meaning a card separate from 16 the EBT card that SNAP recipients currently use, like 17 a debit card, and using the same SNAP EBT card. 18 They said, if you're going to pilot this, 19 really, the easiest, the less costly option, would be 20 to do it with paper, because this wouldn't require a 21 lot of changes to the EBT system. However, there's 22 disadvantages to using paper, and most people adjust 18 1 to that. 2 They didn't want to use paper, for all 3 the reasons that SNAP would move from paper to 4 electronic delivery in the first place, because of 5 potential stigma. That was not at all the preferred 6 way of doing it. 7 We also discussed various options for 8 using a separate EBT card or a retailer gift card. 9 But, generally, people did not prefer that option, 10 because the cost associated with getting behind the 11 card and distributing the cards and replacing the 12 cards. 13 The preferred option by most people we 14 interviewed in our expert panel, was using the 15 existing system, because it would build on the 16 separate checkout system, and may require less time 17 that using a separate card and reduce progress 18 And so the transaction scenario that we 19 discussed in our expert panel, went like this: using 20 the existing EBT cards, participants bring up their 21 food. Most retailers have scanning systems. The 22 items we scan, the cash register systems, separated 19 1 out the target or healthy food purchases from other 2 items. Right now, they just flag the SNAP- or WIC- 3 eligible items. 4 They would separate out and subtotal 5 those. That would go through the system so the 6 participant will be able to swipe their card and 7 enter their payment one time, and the EBT contractor 8 can calculate the incentive amount at the point of 9 sale, saying what the amount should be, which would, 10 in effect, be a discount on the cost of the food. 11 While this option has some advantages, it 12 has disadvantages. The main one is it would require 13 several changes to the EBT system, potentially costly 14 changes. 15 For example, retailers and contractors 16 would need to make modifications to their software to 17 be able to separate out and track healthy food 18 purchases from the other purchases. On page 32, we 19 highlight some of the changes, based on our 20 discussion that we had on what needs to be made, in 21 order to implement that kind of scenario. 22 The final area I wanted to touch on, was 20 1 monitoring and evaluation. Of course, monitoring and 2 evaluation are critical to protecting program 3 integrity and identifying the effects of the program. 4 We heard that many incentives, however 5 you did it, would potentially increase progress. 6 One scenario that we heard about, that you might 7 want to consider, is, what would prevent someone 8 from going into the store and buying all the 9 strawberries and blueberries and getting a discount 10 on those items, and then going and selling the items 11 somewhere else? 12 Considering, you know, it may be very 13 unlikely, but they're things to think about, in 14 terms of how best minimize this. 15 One idea might be to maybe make some sort 16 of limit on the amount of purchases that we can get 17 to discount. 18 Evaluation issues: We're going to have 19 the panel on evaluation. There's a lot of issues 20 that we have to report, that we will discuss further 21 today. 22 How does that confine people to 21 1 comparison groups and random assignment? What 2 outcomes did you look at in the discussion paper? 3 It looked like FNS wants to look at how purchases 4 have changed. 5 They brought up in their discussion 6 paper, the need to look at the whole; how people 7 spend their whole grocery budget, to be able to see 8 a certain substitution effect there. 9 Then data collection. Where do you get 10 that data, whether from retailers or participants is 11 the best way. 12 One of the interesting points that was 13 brought up on our panel, that you might want to 14 consider, is whether researchers would need -- 15 somehow need to isolate this change from other 16 promotions, other promotions or other changes in the 17 pricing of foods that you're looking at, and whether 18 that needs to be considered or not, however one would 19 design the evaluation. 20 Then, of course, the duration of a pilot; 21 how long is it going to take to do this and be able 22 to respond to these incentives? 22 1 So, those are the issues I wanted to 2 bring to you today. I hope that provides some 3 context for your discussions. 4 I'm happy to take questions. I don't 5 know if we want to go right into those. 6 MR. CARLSON: We'll hold our questions 7 till later. 8 MS. ROARK: I'll be here all day. 9 (Applause.) 10 MR. CARLSON: If the first panel members 11 could join me up front, and while they're doing that, 12 I forgot to mention that as you need to leave, if you 13 need assistance in finding a taxi or cab, see Kim 14 here in the back and she'll make arrangements to get 15 one to you. 16 I also wanted to mention that the fact 17 that we're all here today, reflects the hard work of 18 a large team of folks. If I tried to name all of 19 them individually, I'd miss some. My memory isn't 20 that good, but I know who you are, Jessica knows who 21 you are, and you know who you are. 22 I really appreciate all the effort that 23 1 went into this, and the rapid fashion in which you 2 did it. I was going to make some judgment on the 3 success today, but I think I'll withhold that. 4 (Laughter.) 5 MR. CARLSON: The structure for the day, 6 is going to be simple: We're going to have a series 7 of panels focusing on a specific topic, either the 8 initial implications, the evaluation requirements, 9 or system issues that affect the design and 10 implementation and development of the Healthy 11 Incentive Pilot Project. 12 The point of this, if it hasn't been made 13 clear already, is that what Kathy described for us 14 momentarily, was what we know, which, despite the 15 hard effort of the Government Accountability Office, 16 is, frankly, not much. 17 There are a lot of questions that we have 18 in front of us in the design of the pilot and the 19 evaluation of the companies. We start with not 20 quite a blank slate, but pretty close. 21 The goal is to try to describe the issues 22 and the choices that are ahead of the Agency, all 24 1 with the goal of leading us to solicitation for 2 applications to run this demonstration and the 3 contract to provide an independent evaluation. 4 It may sound, as I lead us through this 5 discussion, like we've made some choices already. 6 That's not entirely true. 7 We're open to persuasion on all the 8 critical points and most of the small points. I may 9 sort of shape or share with you, some of our 10 thinking, as the panel progresses, to help keep us on 11 track, if, for no other reason. 12 I am going to address a series of 13 questions to each of the panels, separately. They 14 will be tailored to the expertise on the topics. We 15 hope to work our way through all of them. 16 The ground rules for the discussion are: 17 Everyone is invited to speak. We welcome your 18 opinions. 19 We're not seeking agreement. If you 20 don't agree with your colleagues, feel free to say 21 so. 22 We're not seeking consensus; we're trying 25 1 to explore the options and implications that are 2 ahead of us. 3 There will be time at the end for 4 questions from the audience. If someone has said 5 something that requires additional clarification, if 6 there's a point you think we've missed, we will open 7 the floor to participation from the entire group. 8 I'm going to be, if I have to be, 9 occasionally rude, to keep us on track. There's a 10 limited amount of time for us to work through a lot 11 of material. 12 In order to give everyone an opportunity 13 to speak, I may have to focus our conversation or 14 cut off some points, a little early. It's not that 15 we don't want to hear your opinions, but we want to 16 make sure that everyone has a chance to voice their 17 ideas. 18 I may occasionally try to bring us back 19 on track, if we tend to run off on more interesting 20 side paths, but I will try to restrain myself, so 21 that we actually do explore all the areas. 22 Let's see if there's anything else. I 26 1 think that's the basic set of ground rules. 2 Let me start by introducing the panel 3 members, and let me start by getting the list of 4 them. 5 (Pause.) 6 I warn the audience and I will warn the 7 panel themselves, that we're a first-name agency. 8 If I call on you or ask you for input, by your first 9 name, that's not out of any lack of respect for your 10 eminent qualifications, but that's who we are. I'm 11 Steven, and you are who you are. 12 (Laughter.) 13 MR. CARLSON: I'll walk our way down the 14 table. First we have Elizabeth Pivonka, President 15 and CEO of Produce for Better Health Foundation; 16 Alison Yates -- I followed you around and around, 17 but currently, who is serving as Director at the 18 Beltsville Human Nutrition Research Center, as part 19 of the Beltsville Agricultural Research Center; 20 Debrah Palmer, who is an Associate Extension 21 Specialist and Director of the New Jersey Food Stamp 22 Nutrition Education Program. 27 1 Timothy Mooney is the Director of the 2 Bureau of Supplemental Food Programs in the Division 3 of Nutrition of the New York State Department of 4 Health; Rafael Perez-Escamilla is Professor of 5 Nutritional Science and Public Health at the 6 University of Connecticut. 7 Isobel Contento, the Mary Swartz Rose 8 Professor of Nutrition and Education at Columbia 9 University; Patrick Druhan, who is Director of the 10 Food Resource Center and is also here representing 11 the Association of State Nutrition Network 12 Administrators. 13 Finally, but certainly not least, Sharon 14 Glass, Vice President for Health and Wellness of 15 Catalina Marketing. 16 Welcome to all of you. We're just going 17 to dive right in. The point of this panel is to try 18 to deal with some of the fundamental issues, 19 nutritionally-related issues, and what we're trying 20 to do by encouraging consumption of fruit and 21 vegetables. 22 And I guess most of the questions that I 28 1 have to pose, really do focus in on providing 2 incentives for fruits and vegetables. I will say 3 that while the legislative authorization gives us 4 the ability to consider other healthful foods, we 5 are pretty strongly leaning to focusing the initial 6 effort of this panel, on fruits and vegetables. 7 So let me just start with a question that 8 wasn't on the list that we shared with you up front. 9 (Laughter.) 10 MR. CARLSON: Given what we have to do, 11 does it make sense for us to narrow the scope of the 12 target foods we provide, to narrow it to fruits and 13 vegetables? Then I can use the rest of my 14 questions. 15 Here's the first one for discussion: If 16 we agree that focusing on fruits and vegetables, 17 makes sense, are there some reasons or what are the 18 pros and cons of providing incentives for all fruits 19 and vegetables, as opposed to some fruits and 20 vegetables. 21 Given the nature of fruits and 22 vegetables, do we incentivize corn? It could be the 29 1 form of the fruits and vegetables. Is it important 2 to incentivize fresh as opposed to canned and frozen? 3 And if we are rewarding the purchase of 4 all fruits and vegetables, is it easier to apply 5 rules of thumb to distinguish between what we want 6 to encourage, and things that some might argue are a 7 little less or a little more marginal? That's not to 8 disparage them, but our Whole Foods market is full of 9 frozen fruit bars and a variety of things, but they 10 may or may not deliver the same nutrition. 11 So, the first question is, all fruits and 12 vegetables? Some fruits and vegetables? If so, how 13 do we figure that out? Elizabeth? 14 DR. PIVONKA: I'll start on that one. I 15 think we should actually follow the dietary 16 guidelines for all forms of fruits and vegetables. 17 We had a program, together with the National Cancer 18 Institute, called Fruits and Veggies Matter. 19 This has been -- and we partner with 20 CDC, the National Cancer Institute. Ultimately, we 21 need to support the dietary guidelines for 22 Americans. 30 1 I will say, in our research, as we 2 transition from the fruits and veggies, one of the 3 things we learned from consumers, mothers, in 4 particular, is that they felt guilty that they 5 weren't feeding their families fresh, which was a 6 real surprise to me. 7 I think we need to be careful. Where I 8 think we're going with this, is, should we feed them 9 fresh, versus all forms? I think we need to be 10 careful that we don't limit it. I think we have 11 probably, over the past 15 or 20 years, made some 12 lower-income folks feel disenfranchised, if they 13 aren't feeding fresh, and I think if we stick with 14 just fresh, we may exacerbate that. 15 I recognize there are some down sides to 16 promoting all forms, and we can get into that, but 17 at least as to your initial question, should it be 18 fresh or all forms, I would advocate for the all- 19 forms. 20 MS. GLASS: Sharon Glass with Catalina. 21 We do incentives as our business. Because of that, 22 we respond to the audience. 31 1 We give them a choice of form, whether 2 it's frozen or canned. I think they're going to get 3 the greatest response from consumers when it comes 4 to fresh. We all know it requires a lot of 5 preparation. 6 The audience may not be able to do the 7 prep that's required with fresh. So the canned or 8 frozen option is much quicker than other forms, so I 9 think the broader the scope, the better response 10 we're going to get with this program. 11 DR. PEREZ-ESCAMILLA: For the Healthy 12 Incentives Programs to work, first of all, we need 13 to, whatever information is given, we need to be sure 14 that they need to be accurate, before they're 15 promoted. 16 I'm work mostly with low-income 17 communities that don't have enough access to large 18 supermarkets. Half of them don't have a car, and a 19 lot of the purchases make people think -- well, I 20 could go on. 21 I think that if we try to narrow the 22 target to fruits and vegetables and so on, we may be 32 1 running into issues of access, because I'm not sure 2 that most shops are really equipped to stock a large 3 variety of fruits and vegetables right now. 4 DR. CONTENTO: If we're talking about 5 different forms of fruits and vegetables, I'm 6 wondering about the degree of processing, which is a 7 slightly different issue. 8 I can see fresh, canned, et cetera, for 9 whom we're talking about. I'm assuming that french 10 fries and potato chips -- are they considered? Not 11 at all, right? 12 I didn't know what the USDA definitions 13 were. I don't know whether different groups do or 14 not. 15 Veggie chips and other kinds of things, 16 there may be some way to go by degree of processing, 17 but I don't know how you would present a rule of 18 thumb. I couldn't think of a good rule of thumb, 19 myself, for that, but there should be a distinction 20 in that aspect of fruits and vegetables. 21 MR. CARLSON: That's one of the 22 challenges we're facing. If we broadly incentivize 33 1 all fruits and vegetables, at what point does it no 2 longer become a fruits or vegetable? Deborah? 3 DR. PALMER: First, when I first looked 4 at these questions, they were all very, very 5 difficult. I'm glad they were not on my qualifying 6 exam when I got my doctorate. 7 (Laughter.) 8 DR. PALMER: In any event, I gave these a 9 lot of thought. I think that the issue of 10 overcoming the processed foods, et cetera, could 11 easily be resolved by saying that the only product 12 that could be used, are those either in the whole or 13 cut form, and that have nothing added to them, other 14 than water and some type of preservative, if 15 necessary, and sugar. 16 But I also was very concerned about two 17 things: Product displacement and sustainability. 18 The reason I bring up product 19 displacement -- all of you in WIC are probably 20 familiar with this -- back when we offered either 21 peanut butter or dried beans, everybody took the 22 peanut butter. 34 1 Why did they take the peanut butter? 2 Well, because peanut butter costs more than dried 3 beans. I didn't mean they wouldn't buy dried beans, 4 but it meant that they were going to use the 5 opportunity to get the more expensive product with 6 their WIC coupons. 7 We know that that program is not intended 8 to pay for 100-percent of the month's food, so, 9 therefore, when given the opportunity, you're going 10 to purchase more expensive foods with that part, and 11 then get other foods with your money. 12 I was really thinking about those and 13 thinking about our opportunities for sustainability, 14 and I think we need to think every cleverly about 15 what foods people already buy. Giving a discount on 16 that, is not going to make a lot of sense, because 17 they already do that. 18 When do people eat fruits and vegetables 19 and when don't they? When would you like to see 20 them do it? 21 I think that if you look into the 22 research, you're going to find that it's much more 35 1 common for people to consume fruits and vegetables, 2 particularly vegetables, at dinnertime. So we want 3 to look at those things. 4 They might consume it for breakfast or 5 lunch or snacks, and perhaps put the incentives on 6 those products, because those are the products they 7 don't currently eat, that we would like to see them 8 eat. 9 I think we can go back to the data, so I 10 just did a whole lot of stuff. This is sort of what 11 I thought, after thinking and thinking and thinking 12 about this question. 13 MR. DRUHAN: I'm thinking that minimally- 14 process, might be one of the things you might want to 15 do. I really like the idea of being able to focus on 16 how much of the ingredient you might want to use. 17 I think the idea came up for some kind of 18 percentage. It may be a little too difficult, but I 19 think the idea of minimal processing, is one we 20 should focus on. 21 I do like the idea of fresh fruits and 22 vegetables, if we need to limit the pilot. If 36 1 there's no enough money to do all fruits and 2 vegetables, it may be that doing fresh fruits and 3 vegetables, is something that's relatively easy for 4 retailers to deal with. 5 I sympathize, to a large extent, with the 6 idea that it's hard to get fresh fruits and 7 vegetables into the corner stores in urban areas, 8 however, this program, in and of itself, may create 9 an impetus for a lot of these small stores, to 10 actually get those fruits and vegetables. 11 MR. CARLSON: Could I ask how easy or 12 difficult is it going to be, to comply with minimal 13 processing? I ask this in ignorance. I'm not the 14 principle food shopper in my family, but how much 15 processing goes into producing a can of corn? How 16 much processing goes into producing frozen green 17 beans? 18 MR. DRUHAN: Veggie chips or any of 19 those, don't have a whole lot of vegetables. 20 DR. PIVONKA: Can I answer that? That's 21 something we've lived with for many, many years. 22 This is an example of spaghetti sauce; I mean, 37 1 depending on how you define it, spaghetti sauce 2 might be kicked out, whether it has meat products in 3 it or not, because it might be a little too high 4 sodium. 5 I will tell you that it's a slippery 6 slope. It's very difficult, but we've worked on it 7 together. CDC and Deborah Seymour and her group, 8 together with the USDA and National Cancer Institute 9 and FDA folks have identified that what products we 10 consider fruits and vegetables, they looked at the 11 2005 dietary guidelines, and they interviewed how 12 many people would be eating, if they ate all of them 13 in cans, and they set criteria by which they thought 14 we could carry our logo, and I would encourage you to 15 take a look at that as something that might be 16 promoted within this program. 17 An example: Some berries take a sugar 18 bath, just from a processing bath, to maintain their 19 integrity, but from the label, it looks like they 20 have sugar added, so frozen berries might be thrown 21 out for that reason, and we don't want to throw them 22 out. 38 1 Another example would be frozen green 2 beans. CDC has given a credit to products that have 3 nuts in them, because it's a healthy product. We 4 think frozen green beans is just a great 5 alternative. 6 Prior to this analysis, it wouldn't have 7 been able to carry our logo. I could say we could 8 spend a lot of time talking about this, and we 9 probably don't want to spend a lot of time. 10 You might want to take a look at what CDC 11 has already done. They spent a lot of time doing it. 12 DR. PALMER: I'm really concerned about 13 when we think of only fresh fruits and vegetables. 14 Again, with our audience, looking at sustainability, 15 often, canned and frozen products are less expensive, 16 particularly in inner city areas, than the fresh, so 17 if we want people to get in the habit of doing 18 something and continuing it, I don't think it makes 19 sense to limit it to fresh fruits and vegetables. 20 On top of that, in terms of nutrition, 21 often, frozen products have more nutrition than 22 fresh, because the are frozen immediately after 39 1 picking them, instead of being shipped for days and 2 days. 3 So I don't think it makes any sense at 4 all, to limit it. 5 DR. PEREZ-ESCAMILLA: I also want to add 6 that I do fell that farmers markets have a role to 7 play in initiatives such as this one. In Hartford, 8 we have good number of farmers markets. 9 I think we should also include them as 10 part of this discussion. I also wanted to mention 11 that I know that in the Food Pyramid and so on, 12 fruits and vegetables are in the same group, but it 13 doesn't make a lot of sense for me to talk about 14 incentives, behavior, and choices, when it comes to 15 those. 16 There are different fruits and 17 vegetables, and if we are thinking about, you know, 18 which ones need to be promoted the most, not only in 19 terms of economic incentives, but educationally, my 20 educated guess would be that they are difficult to 21 deal with, especially if they want to consume a lot 22 of the nice vegetables that not a lot of people are 40 1 used to consuming. 2 We need to think deeper about fruits and 3 vegetables. 4 MR. CARLSON: That's interesting. Fruits 5 and vegetables is one word in my vocabulary. 6 DR. CONTENTO: In terms -- I like that 7 idea that people are more willing to eat fruit than 8 vegetables, but I like Deborah's idea of some other 9 sort of visible form of not fruits and vegetables, 10 which will be, you can have it canned or frozen, but 11 it would still be recognizable as a fruit or a 12 vegetable, to get around the degree of processing, 13 another way of thinking about it. 14 DR. YATES: My background is certainly 15 not in consumer behavior at all, but I know that 16 although the issues related to inner cities, are 17 important, in very rural areas, the ability to get 18 whatever you can from the 7-11, which is maybe ten 19 mile away, is how you deal with the fresh food 20 products. 21 And I would assume you've got a lot more 22 highlighting of what's going on in Alaska. There 41 1 are areas in Alaska where we have problems in 2 access. 3 So I guess I'm more concerned about how 4 much additional nutrient quality are going to get by 5 going to fresh, versus coming up with how you want to 6 limit other types of processing on fruits and 7 vegetables. Is there a better tradeoff? 8 I think the other aspect is, if you're 9 expecting people to get the same message, if there 10 are differing requirements or criteria for what's 11 allowed in packaged, versus what's allowed as an 12 incentive, there's a mixed message that people are 13 getting. 14 That has to be resolved, as well. 15 MR. CARLSON: Let me follow up on this 16 issue of messages, because that's one of the things 17 we're struggling with. 18 The way the question is phrased, how 19 important is it to add additional forms of nutrition 20 education or promotion as part of the incentive, in 21 addition to the financial incentive, in order to 22 encourage people to change their behavior? 42 1 Is just providing the incentive, enough, 2 or do we need to promote its use? If so, in what 3 way, and how would that happen? 4 I can see you jumping, Isobel. 5 DR. CONTENTO: As a nutrition educator, 6 obviously, I'm going to say that we need education 7 along with financial incentives, and we need to 8 think of some kind of theoretical framework for 9 that, as well. 10 Perhaps one way to start it, is a more 11 social ecological model, which says you have to come 12 at it from several different angles, which is 13 certainly the incentive to reduce the barriers. A 14 very concrete barrier, is cost, but there are also 15 perceived barriers. 16 There's also motivations to actually want 17 to eat fruits and vegetables, and I think that the 18 GAO report also says something about how. So 19 consumers want to know why it is that it's so 20 important to eat these fruits and vegetables, so I 21 think we have to have other things that go with it. 22 There could be brochures, there could be 43 1 newsletters. 2 Now, obviously, studies show that more 3 intense intervention, such as working with groups 4 and counselling, may be more powerful, but, on the 5 other hand, in this particular kind of situation, it 6 doesn't mean we can't do other kinds of things, as 7 well. 8 It needs to come from different angles. 9 Each one adds something, by itself. Financial 10 incentive are not going to do; education, by it 11 self, is not going to do it. 12 I also think that the financial 13 inventive, is a very extrinsic motivation to reduce 14 costs. What we really want, is for the family to 15 come to like the fruits and vegetables and to 16 eventually move it to be more intrinsically 17 motivated to eat fruits and vegetables or 18 vegetables. 19 They are harder to persuade people to 20 eat. That's where we can do the kinds of things 21 that help people to develop that, and that is for 22 them to get some sense of -- we would have to design 44 1 a program around that -- but some sense of 2 confidence that they feel they can do it, some sense 3 that they have some choice in the matter, and some 4 kind of relatedness that this program somehow is 5 related to them, personally. 6 So, whatever messages are out there, 7 could make them feel that they are related to 8 whoever is promoting it, in a very positive sort of 9 way, so they want to do that. 10 I think we do need to accomplish it so 11 that people will want to do it, as well as having 12 the ability to do it, because the prices will be a 13 little lower. 14 MS. GLASS: Not only that, but we have 150 15 million households in our database, data on purchase 16 behavior and education and messages; that the 17 household did not purchase any vegetables in the past 18 few years. Was it frozen? Was it fresh? Was it 19 canned? 20 We create messages that are specific to 21 each household. We educate them on the reason why 22 we're getting this incentive. 45 1 There would be a messaging campaign first, 2 stating, here's a recipe to how to use vegetables; 3 here's why you should have vegetables in the 4 household to improve your diet. We can identify 5 households who are very unhealthy diet-based, and 6 specifically change the message, based on the fact 7 that that's an unhealthy household, how to improve 8 their health. 9 Here's the messaging campaign, first, and 10 it incentives to instill them or motivate them to 11 actually change their behavior. A lot of our 12 programs have a step-by-step process where we 13 identify the house in our database, then we start 14 with the messaging campaign. 15 If they still have not adhered to the 16 message we are trying to get to them, then we follow 17 up with incentives. The message or education is not 18 hitting home, so let's give them a coupon now, so 19 that they are more willing to make the change. 20 Those are typically the campaigns that we 21 run. 22 DR. PIVONKA: Can I add to that? This 46 1 whole idea that incentivizing people to use their 2 SNAP dollars for healthy foods, is something that's 3 been discussed for a long time. I'm really excited 4 we're having this conversation, but I have to say 5 that when I first thought about this, it wasn't 6 until I read the GAO report, that I realized how 7 complicated it really is. 8 But when I first thought about this, 9 probably in the late '90s, my thought was, Catalina 10 Marketing and all the coupons and the things they 11 can do, was the best way to deliver this program. 12 You can target it to your audience, for example, 13 young mothers. 14 You try to pitch avocados to young 15 mothers, because it's soft for baby food. That's 16 the message you could target, have these based on 17 their demographics. 18 You probably only have to do an education 19 program once. They get it, they figure it out, and 20 you move on to something else. 21 But I think the power of Catalina 22 Marketing, is something that we haven't tapped into 47 1 enough. 2 MR. CARLSON: If I could just say, for 3 those of you who are on the afternoon's panel, one 4 of the things to think about real hard, is how to 5 make the Catalina-like system work in a SNAP/EBT 6 context. 7 As I understand it -- and correct me if 8 I'm wrong -- those programs are largely based on 9 courtesy cards or the preferred customer cards. You 10 sign up for the program. 11 The current EBT system, is largely just a 12 transactional process. We know how much money 13 they're spending, but it's not liked into items they 14 purchase; it's not linked into the scanning system. 15 We know that this is a young mother, but 16 the question for the systems folks, is, what would 17 it take and how difficult might it be to create that 18 link between the transactional system and the store 19 standards, to understand the kind of food choices 20 people make? 21 As one thing on there, in our database, 22 we'll be able to look at every cart, not just the 48 1 shopper carts, but anything -- ATM cards, whatever 2 the number, whatever is scanned through. 3 The same thing goes with SNAP cards. We 4 can identify in our database and we have a good idea 5 of the consumer that used that card, then we can look 6 at their two-year food behavior, identifying whether 7 they buy vegetables at Thanksgiving, or do they only 8 buy at Christmas? 9 We can then track that, and then we can 10 say, okay, in our program, what is the future change 11 of behavior? For as long as you want to, you 12 continue to track those households and say, okay, 13 after our program, we saw a 30-percent increase in 14 vegetables in this SNAP household, and we're tracking 15 them on our system. 16 It's the way they have been targeting for 17 25 years, whether it's a mom with kids, or trying to 18 get them to buy, you know, certain products, you 19 know. 20 We can do the same thing for SNAP cards. 21 MR. CARLSON: I apologize for 22 interjecting myself into this, but that's a good 49 1 point. 2 The other complication that the program 3 faces, that is, if you're linked to EBT cards, you 4 then can't figure out what other foods they have 5 purchased with their out-of-pocket cash. Most 6 families are spending a fair amount of food 7 purchases, out-of-pocket. 8 Have I mentioned that I work in what used 9 to be the Food Stamp Program for 32 years? 10 (Laughter.) 11 MR. CARLSON: It's hard for me to change. 12 Rafael? 13 DR. PEREZ-ESCAMILLA: I'm very much in 14 support of education being included, together with 15 incentive programs, as long as we talk about the 16 SNAP participants. 17 It is instrumental, education, not only 18 information. In other words, hands-on food-label- 19 reading, recipe preparation, menu, budget planning, 20 and so on. 21 We know we have a very vast experience in 22 the country, that has accumulated over ten years, and 50 1 we could definitely learn a lot. I also see another 2 enormous opportunity for education to society at 3 large, from SNAP, because I think this presents an 4 enormous opportunity to send a very powerful message 5 that we deeply care about the quality of the diet 6 that SNAP participants are consuming. 7 I think a lot of people that are not very 8 supportive of SNAP and other nutrition programs, care 9 about the use of the money. They feel very good 10 about that initiative, but I also like the social, 11 ecological model that we think about the recipient, 12 but we also think about the educational opportunity 13 that these present for society at large. 14 MR. MOONEY: As I read through this 15 question, I did some back-of-the-envelope 16 calculations. In reading through the GAO report, 17 they mentioned 26.7 million SNAP recipients, 23 18 million WIC recipients. 19 Then I went back to New York and I 20 looked. We have about two million SNAP recipients 21 in New York, and over 500,000 WIC recipients, about 22 380,000 households, which constitute probably about 51 1 900,000 individuals. 2 Almost one-half of the SNAP recipients, 3 are already receiving quality nutrition education. 4 When I add to that, other programs such as the Child 5 and Adult Care Food Program, our nutrition 6 education, we are reaching many, many of these 7 individuals already, with consistent messaging. 8 Certainly, that's very important. 9 Rafael, you spoke about the total messaging going 10 out to the population, in general. I like to refer 11 to that as sphere of influence that we have, that 12 the messaging that we give to the moms that come 13 into the WIC program or the moms that are 14 participating in child and healthcare food programs, 15 they are taking those messages back to their 16 households and translating them into messaging and 17 education for everyone in the household. 18 So there is a large nutrition component 19 that exists out there, and we need to continue to 20 build upon that. 21 DR. PALMER: I know the question was on 22 education, and, you know, I just keep going over and 52 1 over these things in my head. We've got to start 2 thinking differently and not do the same old thing. 3 We've got to start thinking like a SNAP 4 benefit recipient. I used to be one; I used to be 5 surrounded by other people who were there. 6 You need to think like them and you need 7 to think about how do we increase consumption? 8 Again, I'm concerned about displacement here; I'm 9 concerned that if we lower the price of vegetables, 10 do we have enough money left to buy shrimp? 11 What we're trying to do, is increase the 12 intake of vegetables, increase the intake of fruits. 13 We don't need recipes. There are a 14 gazillion recipes already available. What do people 15 do with recipes? They replace what they were already 16 doing, with this new recipe, should they decide to 17 adopt it. 18 So, that's not what we need to be doing. 19 Again, I'll come back, because I think we have to 20 educate people to do things differently. They need 21 to think differently than they were thinking, if 22 we're going to have them consume more fruits and 53 1 vegetables 2 Again, I think we ought to go back to the 3 research and look at when aren't they getting it, 4 that they could be, then that should be the basis of 5 our message, at least at the point of purchase. 6 If you're saying to a mom, well, if you 7 buy these baby carrots and you put those in your 8 child's lunch, instead of potato chips, we're giving 9 you this incentive and you're going to save money, 10 plus, they're going to be eating a healthier diet. 11 Now, when you put that together, they'll 12 go, oh, so they start thinking differently. But to 13 just give them a recipe, doesn't change how anybody 14 thinks. They are still thinking, well, we can have 15 that for dinner instead of this. 16 You have to come up with ideas, not 17 necessarily mine, but ideas that they're going, 18 okay, what's going to really increase this where it 19 doesn't exist? I don't know how many of you have 20 tried to eat the right kinds of fruits and 21 vegetables yet. 22 I've tried. The only way that you can 54 1 make it happen, is to eat them for snacks and by 2 putting them out for lunch and breakfast. We've got 3 to think that way, or it's not going to happen, and 4 then we need to provide point-of-purchase 5 information, or education to get people thinking 6 that way. 7 MR. CARLSON: The other way to get there, 8 is to have a wife. Patrick? 9 MR. DRUHAN: As a point of disclosure, 10 I'm not a researcher, I deliver a lot of food 11 programs and nutrition education. I'd like to 12 follow up and say it's true that we have a very 13 complex relationship with food. 14 It's not just the economic incentive 15 that's going to be needed to change behavior. In 16 Pennsylvania, we had a proposal in the farm bill 17 last year, that would say, if people buy fruits and 18 vegetables, they get a bonus. 19 So we targeted it at 50 percent. In 20 order to get the 50 percent, they would actually 21 have to change their behavior. 22 That would probably be better in terms of 55 1 an incentive, if it were targeted that way. If you 2 just give people extra money, they're not going to 3 switch over to fruits and vegetables. 4 They know that fruits and vegetables are 5 better for them, but their patterns of behavior and 6 their self image, might be better served, if they 7 buy a much better cut of meat, rather than a fruit 8 or a vegetable. 9 And so the nutritional educational part 10 of it, is really important, so people start to feel 11 that eating fruits and vegetables, is something they 12 want to do, whenever we give them the chance. 13 People really like to talk about what 14 they do. We hand out recipes and it works for some 15 people, but it's really getting people involved. 16 I've seen them engaged, talking about 17 their lifestyles, how they incorporate those foods 18 into their daily lives. It really is a cultural 19 relationship that has to be realized, and it's going 20 to be addressed through nutrition education and not 21 through any kind of incentive. 22 DR. CONTENTO: I'd like to support that. 56 1 We tend to use the words, "nutrition education," as 2 nutrition information, and it's not. 3 It's a good deal more broad than that. 4 It's all the kinds of things you've talked about, 5 but I just want to make sure we think about that. I 6 don't know what other words we want to use, instead 7 of education, but it is -- 8 MR. CARLSON: Let me pursue that just a 9 little bit. The SNAP program is currently serving 10 something like 29 individuals. Half of them are 11 kids, but a large number of them are adults. 12 If you think of these pilots with the 13 goal, ultimately, of going to scale and providing 14 this incentive and it being accessible to all of 15 those households, what does that imply for how we 16 actually deliver and design nutrition education and 17 promotion? 18 There are very few settings in the SNAP 19 program, other than initial certification exposure, 20 where you actually have the family in front of you. 21 Are there ideas of how you can reach 22 large numbers of people, relatively inexpensively, 57 1 and with information that is effective in terms of 2 changing and not only providing information, but 3 also the motivation or changing their behavior? 4 Elizabeth, I don't mean to put you on the spot. 5 DR. PIVONKA: Actually, there's a lot of 6 thought along these lines. 7 McDonald's spends $1.6 billion a year in 8 marketing and advertising. We've been looking at 9 the success of many of these commodity boards, 10 including dairy, beef, and they're successful. 11 They've show that there's a return on 12 investment, if you spend money marketing those 13 products. 14 I don't know if you want to call it 15 nutrition education, marketing, whatever you want to 16 call it, but it's all doing the same thing; it's 17 impacting what a person thinks about their food; 18 it's not just what I do with it and what am I going 19 to serve today; it's how they identify with their 20 food. 21 We need to get to that somehow. I don't 22 know if it's through this incentive program, or, if 58 1 we pull this incentive program together with the 2 FSNE programs and somehow tie those together. They 3 need to complement each other and work together. 4 I know FSNE is intensively focused on 5 their target audience, but there is a piece to this, 6 that if we have a stronger national effort for the 7 general population to increase fruits and 8 vegetables, then it's not just targeted at lower- 9 income; it raises everybody up to, hey, this is 10 cool, this is something I want to do; it's not 11 something I'm doing, just because they're telling me 12 to do it; it's good for me; it's something I want to 13 do. 14 I don't know that that's what we can do 15 here, but we need to recognize that that's a 16 component of all of it. 17 MR. MOONEY: I'll go back to how large is 18 large. When 50 percent of your recipients are 19 receiving consistent messaging, I'd say that's 20 pretty large. 21 Going to the product displacement issue, 22 something that I struggled with three years ago when 59 1 we did a pilot project in New York, I think one way 2 of measuring the effect of product displacement, was 3 by asking individuals, did you buy something that you 4 had never purchased before? 5 When we found that 40 percent of the 6 respondents said, yes, I bought something that I had 7 never eaten before, we didn't displace anything. We 8 introduced them to new foods and new ways of 9 preparing them. 10 So, again, 40 percent is a fairly sizable 11 number. If we had a 40-percent success in all of our 12 public health interventions, we would be 13 congratulating ourselves. 14 MR. CARLSON: Rafael? 15 DR. PEREZ-ESCAMILLA: Can I ask a 16 question to Elizabeth? I've always been very 17 curious; the beef industry, the poultry industry, 18 the dairy industry, as you said, have been very well 19 organized, very successful at marketing. 20 I always wonder, why don't fruits and 21 vegetables? It's a fair question. I asked the same 22 question in 1990, when I started working with this 60 1 issue. Why don't they do that? It's a very good 2 question. 3 They do it at the commodity level. The 4 California Table Grape Commission, the Pear Bureau; 5 there's one on tomato, avocado, a Watermelon Board; 6 they have it at each commodity level. 7 We're trying to justify to the industry, 8 that -- and we haven't been able to do it yet. 9 We're trying to justify to them, why it would make 10 sense that they do it on a national level. 11 Our program is voluntary contributions. 12 Our budget is very tiny. Those are mandatory 13 programs. 14 The growers vote to invest, themselves, 15 so every head of cattle that's sold, a dollar goes 16 back to promotion. Every gallon of milk that's 17 sold, a dollar goes back into promotion, for every 18 load of grapes that's sold, it goes back to grape 19 promotion, but there isn't anything for a national 20 fruits and vegetables promotion. 21 We've had discussions, we're talking 22 about it, but it's not anytime soon, unfortunately. 61 1 What would be nice, is if we had enough funding to 2 show a test of how it would be effective, but that 3 takes a lot of money, too. 4 MR. DRUHAN: We do incentivize dairy and 5 cattle and a number of other products, through 6 subsidies. We lower their price and that lower 7 price increases consumption, but they are being 8 subsidized. 9 DR. CONTENTO: I said, well, maybe we 10 should start thinking about subsidizing fruit and 11 vegetable growing. That might solve this problem 12 rather quickly. 13 I think you are asking whether, because 14 there's the SNAP program as well as this pilot we're 15 talking about, obviously, they need to be integrated 16 with each other or consistent, because you're saying, 17 in fact, in New York State -- I'm from New York and I 18 did not realize it -- reached quite as many people as 19 you said. 20 Obviously, we need to link them in some 21 way, so the SNAP program could, in fact, tell people 22 about this pilot. The pilot could also refer people 62 1 to other activities that are going on with FSNE or 2 SNAP. 3 DR. PALMER: I want to answer your 4 question on how can we do it. It's very difficult 5 to reach all the SNAP clients with healthy nutrition 6 education. 7 I run the program in New Jersey, but to 8 do intensive education, costs way more money than we 9 have, way more than we'll ever see, if we're going to 10 get to everyone. 11 Our most successful work, I think -- 12 well, again, we can't measure it quite yet -- but in 13 terms of reaching more people, our most success has 14 been that we've trained the people who work at the 15 SNAP office, to actually give out some information 16 and state very, very targeted messages with regard to 17 our marketing campaign. 18 I don't know how effective it is in terms 19 of behavior changes, but I do know that's our most 20 effective means of reaching the largest number of 21 people within the program. 22 With regard to providing education, I can 63 1 tell you that shelf-tuckers are not the answer. 2 We've done that. We've learned that grocery stores 3 are noncompliant, they move their shelving around 4 all the time, there's no room on the shelf, anyway. 5 WIC is probably the most successful, 6 because they have these great big white and black 7 letters, and it's only in the cereal aisle, but, you 8 know, it's too difficult. That doesn't work. 9 We know a lot of things that don't work. 10 The things we've seen that do work, with the 11 interventions I've done, is to catch the attention 12 of children, because their parents get exhausted by 13 them and are happy that their attention has been 14 captured. 15 I would propose that the sorts of things 16 that might work, are to have the continuous loop 17 television -- looks at television, too -- in the 18 produce section, and to, like McDonald's have a 19 character or group of characters in flashy colors, 20 that attract children's attention, to, again, open 21 up their minds to using the fruits and vegetables as 22 snacks, in different ways like that. 64 1 I think that if the kids want it, the 2 parents buy it. That's been pretty well borne out 3 in the literature. 4 MR. CARLSON: Sharon, I don't mean to put 5 you on the spot, but you talked earlier a little bit 6 about the marketing program that Catalina runs, 7 triggered by purchase behaviors and so forth. 8 Do you have a sense or evidence that 9 those are actually fairly effective, that people do 10 respond to the coupons or other incentives? 11 MS. GLASS: Our average response when we 12 do campaigns -- it's not all recipes, but we do 13 things about targeting people that have heart 14 conditions. We give them message campaigns about 15 one aspirin a day, will improve on your health, and 16 we're pretty up front with a lot of these messages. 17 We know that if these people have kids in 18 the house, you can be a little bit more practical 19 with the message. What have you done today to 20 improve the health of your child? 21 So we know households that have children, 22 and we have 35 million in our database, so we target 65 1 them with programs, continually, about whether it's 2 Quaker Oats trying to get oatmeal in the household 3 for a healthy breakfast to start the day, or it's 4 Kraft trying to do something about cheese in the 5 household, a healthy snack for lunch, we do a lot of 6 those things. 7 The main thing would be, this is a trial 8 program. We do retention analysis on whether or not 9 these consumers continue to adopt this performance, 10 so we'll track them for a year. 11 Over the course of the year, it did pay 12 out for this company. They pay out for a targeted 13 consumer and give them $2 a day and we see an 14 average of 40 to 55 percent retention, because we're 15 targeting the household or identifying the household 16 who represents your core group that you want to go 17 after, not really going after the masses, who may not 18 have kids. 19 If you have kids, it really helps the 20 response. 21 DR. PEREZ-ESCAMILLA: I do believe that 22 to reach out with a new initiative such as this one, 66 1 to millions and millions of program recipients, 2 social marketing is a very important tool to use. 3 But it has to be culturally tailored. It 4 is social marketing. Our experience with our 5 Hispanic SNAP ED programs in Connecticut, we've done 6 five social marketing campaigns. They are great for 7 awareness; they are great for reinforcement of 8 messages in the environment, but by itself, I don't 9 think it would be enough. 10 But as we have been saying, there may be 11 enough initiative, enough points of opportunity, 12 already on the ground, and what we may need, is 13 better coordination among them, but not necessarily 14 to come up with a whole new program, education 15 initiative. 16 MR. CARLSON: Allison? 17 DR. YATES: The one thing that reminded 18 me, is, ARS currently has a major project with all 19 the human nutrition research centers, in trying to 20 identify facilitators to consuming dietary 21 guidelines, as well as barriers to dietary 22 guidelines, using the social methodology, and 67 1 actually using fifth graders, and then their 2 caregivers, and going into kind of a focus group 3 methodology, normative group technique, to really 4 find, by various ethnic groups, as well. 5 I'm not sure of the extent of the social 6 aspects, in terms of economics. This project is to 7 be going on for the next two years, but it hopes to 8 be able to get at what are some of the things that, 9 culturally, are constrained purchases, using each of 10 the dietary guidelines. 11 Although the dietary guidelines will 12 probably change, it will be a way to approach, get 13 some of this information. It's not going to be too 14 helpful for what you need right now, but I think it 15 will add to the information available. 16 MR. CARLSON: Let me shift gears a little 17 bit and point out the obvious. Nothing we do, is 18 going to likely immediately change people's 19 behavior. That raises a question of how long do w 20 need to run whatever it is we decide to run, in 21 order to expect to see it again? 22 It has two implications: The important 68 1 point is, one, if we're going to succeed, we'd like 2 to be able to maximize our chances for doing that. 3 But to run the pilot too long, means we start 4 carrying a fair amount of costs, both for the 5 operation and for the evaluation expectation, as 6 well. 7 At some point, we're going to have to 8 balance the need to run it long enough to see an 9 effect, but not so long that we can't afford it. Do 10 any of you have a sort of sense in your experience, 11 of what a minimum desirable time for operating a 12 pilot of this nature, might be, and whether we're 13 likely to see some consequences? 14 DR. PIVONKA: I don't know if I have a 15 full answer to this, but I know, from a marketing 16 and advertising perspective, twelve weeks is always 17 a good number. 18 If we're doing any advertising or 19 marketing, that 12 weeks may go off for six weeks, 20 back on for six weeks. The other component to this, 21 is that, at least with marketing and advertising, 22 separate and apart from incentives, once you stop, 69 1 ultimately, it does go down. 2 It's one of the kinds of things that you 3 need to keep after it. You may have to start 4 stronger, initially, to get the message out there, 5 then reduce it after that. But you can't stop. 6 Now, for your test, 12 weeks might be a 7 good number. 8 MS. GLASS: We typically run, minimally, 9 12 weeks. It's long enough, but the point is 10 building compliance and frequency with that message. 11 It's always part of a program. 12 We run programs for the entire year. At 13 the beginning of the year, we do an incentive- 14 based; the rest of the year, we do the message. 15 Sixty-six percent of consumers who buy vitamins in 16 January, don't buy them again the rest of the year. 17 If they are incented to buy vitamins, you 18 know, half of them -- the incentives are still, how 19 do you maintain that constant awareness and make sure 20 they know the benefits of what they're doing? 21 DR. PALMER: I think that when we're 22 looking at how long, you have to look at public 70 1 health, as opposed to single-product marketing. 2 Twelve weeks, is probably appropriate, when you're 3 trying to market a single behavior with a single 4 product, but I think we're looking at multiple 5 behaviors and a multitude of products. 6 So, it's a very different situation. 7 The other thing is, you have to consider 8 the depth of influence you want to have. If you 9 look at diffusion of innovation, you know, you're 10 going to see some people switch over fairly easily, 11 but for the majority, it takes time. 12 You can look at data, even for a single 13 behavior, although it is somewhat addictive, from 14 the cigarette changes, it's years and years. I think 15 we have to look at a much longer time period than 16 would normally be used in marketing, again, because 17 we're not looking at a single product or single 18 behavior. 19 DR. PEREZ-ESCAMILLA: I think what is 20 important here, is the monetary incentive. We 21 really don't know, just with nutrition education, if 22 there is a fiscal incentive. In our targeted 71 1 community, for example, $10 per month of fiscal 2 incentive, can really motivate people to do a lot of 3 things. 4 DR. CONTENTO: I'm not sure that we have 5 the data to show how long nutrition education needs 6 to be, but I think it's fairly substantial. Maybe 7 Tom Baranowski knows a little bit more about what 8 the time would be. 9 Something I saw many years ago, was from 10 an evaluation of school health measures on health 11 education. There, they talked about 30 or 40 hours 12 of education to get the behavior change part of it. 13 If you spread that out, we're talking 14 about a fairly sizeable amount of time. Of course, 15 we're not working with foods, necessarily, but when 16 it comes from multiple places, multiple sources of 17 education, and maybe it all adds up to a fair 18 amount, coming from SNAP, coming from the grocery 19 store, coming from other sources, as well. 20 MS. GLASS: I agree that different groups 21 are going to respond more adequately and quicker than 22 others. In a lot of campaigns we do, we track what 72 1 they do in the first four weeks. 2 We may get the message that the they 3 still have not adopted. It might take 26 weeks to 4 get to what a group of households are doing, so 5 these costs are based on how they're adapting to the 6 program. 7 One group might need three campaigns and 8 incentives; another group might need it one time, 9 because they knew all along, they've seen all the 10 other education out in the marketplace about the 11 importance of vegetables, but they need a little bit 12 of a push. 13 For those people, you may not need to 14 give them as much money to do it. You can track the 15 group's behavior, as one adapts, household-by- 16 household, and segment them differently, as you go 17 forward, so it saves money. 18 MR. DRUHAN: I'm going to say something 19 as an advocate. I don't believe it should be a 20 pilot program. The pilot should be how we run this, 21 going into the future, and have it be a permanent 22 program. 73 1 Having said that, to arrive at what 2 economic incentives, we should have multiple pilots 3 going on in different areas. That would be short- 4 term. 5 You could target new people coming onto 6 the program at the time they receive their benefits, 7 and then it may run just for a few weeks. Then you 8 can say 50 percent consensus or 30 percent 9 consensus, or whatever. 10 There you define the marginal utilities 11 of the program. I think that for behavior change, 12 and to make a lasting effect, you may need economic 13 incentives, but I think you can get a proper level of 14 economic incentive, by devising the program that 15 way. 16 There is a lot of valuable information 17 going forward. 18 DR. PALMER: For those of you who aren't 19 in this business, I just want to let you know that 20 if we run a pilot, for instance, where we reach new 21 beneficiaries of SNAP benefits, and we're doing a 22 good deal of education in the SNAP offices, everyone 74 1 who isn't in this business, might think, oh, you 2 know, if it works, oh, good; then it will be 3 sustained through SNAP ed. 4 I just want to make sure that everyone 5 who is sitting here, knows that there is no way you 6 can do a large-based program that's run through 7 SNAP ed, with the current funding structure. 8 You use resources, going in, but there is 9 no one-on-one match when you're working with a 10 federal program. Just keep in mind, that will not 11 be a sustainable venue should we do it with the 12 pilot. We would have to have other funds that work 13 though health ed or the people who work in the SNAP 14 offices, would have to be trained to do these 15 things. 16 DR. CONTENTO: Maybe this is an 17 opportunity for you to find out what different 18 lengths of time there would be, and maybe get some 19 sense of what is the length of time that would at 20 least have measurable effects. 21 MR. CARLSON: I was going to try to pull 22 this together. A couple of points were made 75 1 together by combining this issue of how long you 2 have to run, with this notion of what do you have to 3 do to sustain an effect. 4 One could have a program, pilot of 5 otherwise, in place. 6 Do we provide the incentive to low- 7 income families forever, in order to sustain the 8 change in behavior? Do we do this for a six-weeks- 9 on/six-weeks-off kind of notion? Do we do that 10 forever, in order to sustain it? Are there other 11 things one could imagine or other ways one could 12 imagine of designing the operation with the 13 incentive in such a way that it would try to ensure 14 that the effects are more permanent? Rafael? 15 DR. PEREZ-ESCAMILLA: Can I ask the 16 question back to you, what is the thinking of SNAP? 17 Is incentive needed, because fruits and vegetables 18 are relatively expensive? Or is it needed, because 19 people generally don't like them and you need to 20 motivate them to consume them? What is the thinking 21 of your office, as to why you need incentives? 22 MR. CARLSON: We need the incentives, 76 1 because Congress directed us to fund the pilot. 2 (Laughter.) 3 MR. CARLSON: That's an evasive answer. 4 I think we're really interested to understand what's 5 going on here. Frankly, I think that if you talk to 6 four people, you'll get five different answers as to 7 why consumption and purchases of fruits and 8 vegetables are low. It can't be entirely price, I 9 would argue, because there aren't many of us in this 10 room, I would estimate, that are consuming nine 11 servings of fruits and vegetables a day and we can 12 certainly afford it. 13 It can't entirely be access to the 14 vegetables, because 80 percent of our redemptions in 15 the SNAP program, are going to supermarkets and 16 grocery stores who probably have a reasonable 17 variety of fruits and vegetables. 18 It can't be entirely because people don't 19 have awareness or understanding of the importance of 20 fruits and vegetables; the message has been out there 21 a lot. 22 Timothy has pointed out already that 77 1 there's a core of us that have been this in society 2 for a long time. So where does that leave us? 3 I refuse to accept any more questions on 4 this point. 5 (Laughter.) 6 DR. PEREZ-ESCAMILLA: Those are studies. 7 So, the answer to this question, I think, is, it's 8 initial price, and the U.S. Government decides to 9 subsidize these products, as they do with to her 10 commodities, then maybe the incentive will not be 11 needed, in the not too distant future. 12 If the issue is that people don't like 13 enough vegetables, but if we're discovering that if 14 you pay them to eat them, they will eat them, then 15 maybe you want to think about keeping it for the 16 longer term. 17 MR. CARLSON: We've learned from a 18 school-based snack program -- 19 DR. PALMER: Following up on what Rafael 20 just said, much like when you go to the drug store, 21 they'll show you what was eligible for insurance 22 reimbursement and what wasn't. If there is a way to 78 1 flag that through the scanning system, somehow, and 2 only offer the incentive, if you met this threshold 3 of expenditures on fruits and vegetables, I think 4 that would be effective in sustaining it. 5 DR. CONTENTO: as I understand it, the 6 Food Stamp Program, in the past, was certainly a 7 food program, and you could spend your dollars on 8 any food, pretty much, with some limitations. 9 I think it is an exciting and new measure 10 of what you're saying, is that we need to rethink how 11 do we help people rethink how to use those dollars 12 and spend them on healthier foods? 13 Of course, we don't want to state, well, 14 Food Stamps will only be limited to certain kinds of 15 food we would want to use, so we're using an 16 encouragement, a persuasive approach, as opposed to 17 a regulatory approach, that they can only spend it 18 on that. 19 That's what I interpret what we're trying 20 to do. It is, how do we get people voluntarily to 21 want to eat some of the healthier foods that are in 22 the supermarket? 79 1 MR. MOONEY: Just one comment: Earlier 2 on, we identified the issue of access and the 3 availability of food. In the New York City area, 4 with tens of thousands of stores, food deserts 5 exist. 6 The accessibility and availability of 7 fresh produce, fresh fruits and vegetables, is 8 lacking in some areas that are principally populated 9 by lower-income populations. If we are looking to 10 sustain the ability of these individuals to make the 11 purchases, we just be looking at increasing the 12 accessibility and availability of these foods. 13 That is not necessarily only accomplished 14 through empowering, financially, individuals to make 15 a purchase, but also reaching out to the retailer 16 community. 17 MR. CARLSON: Let me shift gears again 18 just a little bit. One of the important questions 19 that we have to wrestle with, before we can use the 20 evaluation, ultimately, to provide us with 21 information, is, how large an incentive do we need 22 to provide, in order to try to affect individuals' 80 1 behavior? 2 In all likelihood, at the risk of tying 3 our hands a bit, we'll end up testing a variety. I 4 don't know that anyone knows what the right answer 5 is, but even if you test a variety of levels, you've 6 got to start somewhere. 7 Do we start small? Do we start big? Do 8 we have a wide range? Do we have a narrow range? 9 Deborah? 10 DR. PALMER: I doubt that any of us have 11 any idea. I think you'll have to find that out 12 during your pilot. 13 MR. MOONEY: I think, again, going back 14 to a pilot project we operated in New York, if you 15 give someone $5 worth of benefits, approximately 80 16 percent of the individuals will use those benefits. 17 And of that $5 that you give them, 18 approximately 90 percent of that value, will be 19 expended, so if you do the math and look at the 20 number of individuals or recipients that you would 21 provide this benefit to, and what the expected 22 utilization rate is of that benefit, and what the 81 1 total dollar amount that they would expend of that 2 benefit, that will give you a sense of how many 3 people can I serve. 4 We used $5. The new WIC food package has 5 set incremental amounts, based on the category, $6 or 6 $8, or $10. There is some basis, some medical and 7 nutritional justification, for arriving at those 8 values, so, again, looking at consistent messaging, 9 looking at the limited funding that we have available 10 and the desire to reach as many individuals as we 11 possibly can, we do have better than just anecdotal 12 information. 13 We have some, I believe, well founded 14 bases for calculating the amount, as well as what it 15 will ultimately cost us, in providing the benefit to 16 those individuals. 17 DR. PALMER: I don't know if we know, in 18 terms of trying to increase the number of servings or 19 the amount of products people are eating. I hate to 20 say this out loud, but I will, I guess; one of the 21 problems we've always had in the SNAP, formerly known 22 as the Food Stamp Program, was that benefits are 82 1 based on how many people are in the household, as 2 opposed to what their ages are. 3 And so I've long held that if you have - 4 - and I talk to participants every day and I know 5 this is true -- if you have three kids under age 6 five, and they're in Head Start and they're getting 7 WIC, you're rolling in food stamps or SNAP benefits. 8 If you have three teenagers, you're in 9 trouble. And that's long been a problem. When we 10 look at increasing fruit and vegetable intake, we 11 also have to look at age, gender, and meeting 12 sufficient levels of what we'd like to see them 13 consuming. 14 I'm just saying that this is an 15 incredibly intricate mathematical formula that we're 16 going to have to look at, if we're going to actually 17 increase servings. 18 MR. CARLSON: To put the question another 19 way, one of the challenges we're facing, is that we 20 know, in the WIC farmers market nutrition program and 21 the seniors farmers market nutrition programs, that 22 providing a voucher of $10 or $15 a month, does get 83 1 spent. 2 The real question is whether it gets 3 spent in such a way that it actually increases total 4 consumption of fruits and vegetables, or, as has been 5 pointed out, is it simply displacing purchases they 6 would have made otherwise? 7 The flip side of that, is, well, how much 8 change can you really expect for $10 or $15 a month? 9 My friends at the Economic Research Service, have 10 done a little research, based on the data that 11 exists, and I don't have the numbers readily at the 12 tim of my tongue. Some of them are here and maybe 13 we'll answer more generally, but it does suggest that 14 reducing the price of fruits and vegetables by 15 something like ten or 20 percent, might get you an 16 increase of a fraction of a serving. 17 The question ultimately is, is that 18 enough of a change to be excited about? ANy 19 positive change, presumably, is for the better, but 20 if we're going from three servings a day to 3.2 21 servings a day, would we call that a success? 22 Would we need a larger incentive to get a 84 1 bigger bang for that buck, and can we afford a 2 larger one? 3 DR. PIVONKA: I don't think you can 4 underestimate the impact of them just getting the 5 incentive. I keep referring back to the Catalina 6 Marketing and just having the message in front of 7 them, as a reminder, that alone can help steer them, 8 whether they use the coupon or not. 9 I think that constant reminder is 10 helpful. 11 MS. GLASS: It depends, really, on the 12 end design of the program. If it's a threshold 13 program, the incentive would be different, because 14 you're tracking the household and saying, if they 15 purchase, within a month, $20 worth of vegetables, 16 we give them a $5 voucher. 17 Then there is the education. We tell 18 them, you're $10 away from exceeding your $5, and 19 that would be like a reminder campaign that there's 20 a threshold and they continue to go after that at 21 the end of the day. 22 It will be the actual product that they 85 1 buy that day; if they buy $3 of actual vegetables 2 that day, would it be different, if it's different 3 from the actual retail value of the products we're 4 including in that program? Typically, we say 40 to 5 50 percent of the average retail price of the 6 product that my coupon is good on, would be enough 7 for the incentive. 8 If we give $10 instead of a $2 incentive, 9 do they buy $4 worth of vegetables that day or 10 something? I really think it's a decision we have to 11 make when we're ready to design a program. We're 12 designing the threshold, and then they have to reach 13 $20 and actually buy $4 that day. 14 DR. PEREZ-ESCAMILLA: The reality is that 15 you have to operate within constraints in terms of 16 the budget that you have. I think your office has 17 probably a fairly broad idea of how much this could 18 be. 19 You may not have a lot of choice in terms 20 of how far you can go with it, but you have a choice 21 in terms of the outcomes on which you choose to judge 22 this. 86 1 I know this is a subject for the 2 afternoon, but I think it's very important that you 3 don't set the bar too high when you're trying to 4 start a program like this, because if you go 5 directly to dietary intake behavior, versus, first 6 of all, demonstrating the food process evaluation, 7 that you have a program, and, secondly, to show 8 that, you know, sales are going up for the 9 commodities you're trying to promote. 10 To me, that would be the way to start 11 with the first wave of research effort. If that 12 looks promising, then go segue and take it to 13 another level. 14 If I remember correctly, the margin to 15 consume for food stamps, is around 30 cents, and for 16 every dollar in food stamps, families spend around 30 17 cents. 18 MR. CARLSON: The estimates range from 25 19 to 45. 20 DR. PEREZ-ESCAMILLA: For fruits and 21 vegetables, I don't know what the answer is, but we 22 might be really looking at a fairly small amount 87 1 here. 2 MR. DRUHAN: The program has just 3 reviewed the cost of fruits and vegetables, and it 4 may even have a perverse effect on increasing 5 consumption of things we don't want them to consume, 6 because you're going to have a substitution effect. 7 You basically are getting income, and 8 that means they're going to have more money to spend 9 on other commodities. So the fact that they're 10 getting cheaper fruits and vegetables, doesn't 11 translate into buying more fruits and vegetables, if 12 you want them to consume more fruits and vegetables. 13 MR. CARLSON: I can't resist observing 14 that Rafael, in his remarks, has pulled the scab off 15 an ongoing debate within our office, whether we 16 should assume that seeing an increase in purchases, 17 is sufficient, or whether we actually need a change 18 in consumption behavior, as well. 19 I'd be interested in where you all fall 20 out on that debate. If you want to know what side 21 I'm on, I'll let you know. 22 (Laughter.) 88 1 DR. PALMER: I think purchase, in terms 2 of canned and frozen, make sense. I'm not so sure 3 about fresh. Fresh often sits in the bottom drawer. 4 DR. PIVONKA: That's a valid point. 5 Actually, we asked that question a lot in our 6 office, too. What is the measure of success? Is it 7 sales or consumption? 8 We look at sales as the more immediate 9 measure and that consumption is the more long-term 10 measure. 11 We usually refer to the government data 12 set or NPD, we refer to a lot. 13 DR. CONTENTO: Are there studies that 14 would link what percentage of food of the studies -- 15 maybe not just fruits and vegetables, but how much 16 people tend to throw away of the food they buy, maybe 17 10 to 20 percent? 18 Could you not use that as a translation? 19 If they purchase more, it's likely they will end up 20 consuming more, unless they tend to throw away those 21 foods more than other foods they might have. 22 DR. PIVONKA: I don't know of any data. 89 1 We have not seen anything other than the 2 disappearance data, where the USDA disappearance 3 data accounts for waste. I don't know if they 4 account for waste in the bottom drawer, but they've 5 accounted for waste along the system, shrinkage at 6 the retail level. 7 I don't know if somebody can answer that. 8 DR. CONTENTO: Years ago, a study in 9 Arizona -- I haven't kept up with it, but people 10 actually looked at the garbage that was thrown out 11 of individual households. They came up with some 12 number like 15 percent that got purchased, brought 13 into the house, but was not used and thrown away. 14 But there should be estimates of what 15 percentage tends to ge thrown away, so that you 16 could get some estimate of consumption from purchase 17 data. 18 DR. PIVONKA: To add to this, one of the 19 things that I can't answer, is, we look at sales 20 data and we hear that fruit and vegetable sales are 21 going up across the board. We look at consumption 22 data, and it's saying pretty flat. 90 1 We look at USDA's disappearance data, 2 and it's going up per capita across the board, so we 3 can't figure it out. There's some discrepancy 4 there. 5 MR. CARLSON: This is the last set of 6 questions I want to pose to the panel, then we'll 7 open it up to the audience, as well: One of the 8 challenges we're facing in the internal 9 conversations we're having, is, how do we let pilot 10 participants know that they are part of a pilot? 11 How do they know there's an incentive 12 there, that they could take advantage of? Are there 13 any potential barriers that are participants are 14 going to likely perceive, or actually face, as we try 15 to enroll them in this pilot, or otherwise? 16 Is there anything we can do about any of 17 that? Is this just an obvious case where whatever 18 we put out there, the demand will be so great that 19 they'll find us and we won't have to find them? 20 Rafael? 21 DR. PEREZ-ESCAMILLA: You know, my 22 opinion -- and I don't have any basis for saying 91 1 this, but in that community where I work, I think 2 the fiscal incentive would be very, very powerful. 3 All you need to do, is to tell one person and they 4 will tell everybody else. 5 (Laughter.) 6 DR. PEREZ-ESCAMILLA: But in terms of 7 barriers, it depends on what the intervention would 8 be. If it's going to change the delivery of the 9 benefit, if it's going to change the way they usually 10 receive their benefits, and they don't understand it 11 well, in terms of giving it more, that could be a 12 barrier. 13 That's something good formative research 14 could be designed that could make this worthwhile 15 for the communities. 16 MR. DRUHAN: You don't want them to feel 17 special. Let them know they're lucky. If you make 18 them feel that they're really being monitored, 19 that's where they feel that there's bias. 20 DR. PIVONKA: I would advocate not 21 telling them at all, because then they might expect 22 it on a regular basis. I think that's why this 92 1 challenge to work this thing is so interesting, 2 because you can target people with different 3 messages, using different amounts, depending on where 4 they are. 5 That's why the Catalina method is so 6 interesting. This is not equal for everybody that's 7 part of the SNAP program, but then there is not this 8 expectation, there's not this automatic shift from 9 one thing to another. It's targeted marketing for 10 their need. 11 MR. DRUHAN: As long as they don't -- 12 DR. PIVONKA: All they need to know is 13 that when I purchase this, I get a ticket at the end 14 and I get a dollar off something next week when I 15 come back. I have to come back to this store, too. 16 We don't need to know that; it's just 17 that there's money now available to give incentives. 18 MR. CARLSON: I think one of the 19 differences, may well be that SNAP is an entitlement 20 program and it is a federal program, as opposed to a 21 private marketing program, so I think there is, in 22 general, an expectation that this will be treated 93 1 fairly, equitably, and equally. 2 On the other hand, it is a pilot; it is 3 an opportunity to try things that might be a little 4 different from the way the program normally 5 operates. 6 MS. GLASS: The program can be very fair 7 and equitable to everybody. It's just a matter that 8 they meet the requirements to get the incentives. 9 I don't think that it's a bad thing that 10 they do talk among each other, that they say, if you 11 buy $20 worth of vegetables this month, you'll get 12 an incentive, did you know about it? 13 MR. CARLSON: And if you buy $50, you'll 14 get more. 15 MS. GLASS: Exactly. So I think you can 16 have exactly the same program, make it equitable, 17 and communicate the benefits of the program. It will 18 not necessarilyl require additional messaging to get 19 more people onboard. 20 DR. PALMER: I feel like a participant in 21 a presidential debate when I say this, because, 22 before I answer that question, I'd like to answer 94 1 another one. 2 (Laughter.) 3 DR. PALMER: On the last question, you 4 had a second half that we didn't go over, and that 5 was in regards to the delivery of incentives. I 6 just wanted to make the point that our target 7 audience, tends towards a paradigm of immediate 8 gratification. 9 But, okay, now I'll try to answer the 10 question that was actually the question at hand. I 11 honestly think that caseworkers are the best way to 12 let people know about it. 13 I 100-percent agree with Rafael with 14 regard to urban areas. A few people find out 15 anything, and everybody knows. 16 This is why fraudulent use displacement 17 to get steaks, et cetera, happens so much. Word of 18 mouth, it just travels like lightening. 19 It's not the same in rural areas. Before 20 I came to New Jersey, I was in Minnesota, where 21 people were very spread out, outside of the Twin 22 Cities. In those areas, they really are, you know, 95 1 dependent on people like caseworkers and social 2 service workers they work with, so I think, to go 3 through the SNAP caseworkers, in addition to, you 4 know, WIC means another employment mandate, but 5 through WIC programs and other programs that we 6 already have, where we do reach a large number of 7 beneficiaries, I think that's enough. 8 MR. MOONEY: Steve, you have enough money 9 to give each SNAP participant, 75 cents, so when we 10 talk about -- 11 MR. CARLSON: Actually not that much. We 12 need to evaluate that. 13 MR. MOONEY: That's right. 14 (Laughter.) 15 MR. MOONEY: So now we're down to 50 16 cents. 17 MR. CARLSON: Maybe ten. 18 MR. MOONEY: We have lots of evaluators, 19 then. So, clearly, this pilot is not for everyone; 20 this pilot is going to be very, very focused. 21 When I first received the invitation and 22 I read through the legislation, I said, oh, $20 96 1 million, okay, I'll take that, just send that to me. 2 I will conduct this pilot for you, and you can 3 evaluate it, because the amount of funding is so 4 limited that it has to be conducted in a very 5 controlled environment. 6 Yes, it is wildfire. Word gets out, and 7 the expectation is there and the annualization of 8 continuing to provide this benefit, you will feel 9 backlash in those local offices, when individuals 10 come in and say, what happened to that benefit that 11 I got last month? What happened to the benefit that 12 my neighbor received? 13 Why? Because I am swimming in food 14 stamps, because i have three individuals on the WIC 15 program, don't I receive this benefit? 16 And I say that there is a disparity in us 17 attempting to improve the nutritional status of that 18 family with three teenagers, because of the lack of 19 other programs that may be available for that 20 population. 21 So, promoting this, is going to have to 22 be done in a very, very controlled environment. 97 1 MR. DRUHAN: You'll get a little blowback 2 on the farmers market, in addition to the Farmers 3 Market Nutrition Program. Everybody else gives an 4 incentive for people to use SNAP at the farmers 5 market, and we do get people coming by the table and 6 saying that, and that get rather angry about it, as 7 well. 8 One other thing, $20 million would 9 probably fit just as well with Montgomery County in 10 Pennsylvania. 11 (Laughter.) 12 MR. CARLSON: There's a sign-up sheet at 13 the door. 14 (Laughter.) 15 DR. YATES: Since we're talking about the 16 small amount of money that's available, one of the 17 questions you didn't really ask, that would be nice, 18 that I'd like to kind of focus on, is fruits and 19 vegetables or fruits or vegetables, because at least 20 if the emphasis of this is to deal with the obesity 21 problem, at least some of the data seems to support 22 increased vegetable consumption being related much 98 1 more closely to lower body weight. 2 So, while there are a lot of good -- and 3 I've got scientists that work all day on trying to 4 see the next phytonutrient in fruit, I really wonder 5 whether focusing on vegetable consumption, where you 6 get at not just related to preparation, ease of use 7 and so on, that's just not there to the same extent 8 with vegetables, as with fruit. 9 MR. CARLSON: That's a fair point. 10 fruits and vegetables is one word, as I said 11 earlier, and I'd like to put hyphens in there, at 12 least. 13 So, there's a whole audience out there. 14 This is your opportunity to ask questions that I've 15 missed. If you want to speak, we're going to ask 16 you to step up to the center aisle, speak into the 17 microphone, and, for the benefit of the transcriber, 18 if you'd please identify yourself, with your name and 19 affiliation? 20 DR. KIRLIN: John Kirlin, formerly with 21 the USDA and with the Economic Research Service. 22 I'd like to make a couple of points that I heard 99 1 briefly mentioned, but I think they need to be 2 reinforced. 3 This is a pilot. Hopefully, it will be 4 more permanent. The Agency needs to keep in mind, 5 what they're thinking of in terms of that permanent 6 program or a permanent change in the staff. 7 They ought to be thinking about what they 8 ought to look at during the pilot, whether it's 9 fruits and vegetables or just vegetables. 10 In terms of thinking about the fact that 11 you have to keep an eye down the line, in terms of 12 what you're looking at, in the long-term, the other 13 point being that a lot of the issues which we have 14 raised this morning, intertwine with some of the 15 issues we will discuss this afternoon. 16 That's the evaluation design and the 17 system design and what can you do or can't you do. 18 One thing I'm very interested in, is 19 whether it's communities that get assigned or 20 individuals get assigned. 21 But, anyway, those -- you have to keep 22 those notions in mind, as you think about all the 100 1 questions that we deal with today. 2 DR. TOWNSEND: Marilyn Townsend, 3 University of California at Davis. 4 I'd like to share with you, briefly, an 5 intervention we did that might answer some of the 6 questions that were raised. 7 This was a four-pronged collaboration 8 between UC Davis, our Physics Program, Catalina 9 Marketing, and a chain, a supermarket chain in 10 California. 11 This particular supermarket chain had 128 12 stores. We decided to target 28 of them that were in 13 lower-income communities. 14 This was for feasibility, to see if we 15 could all get along and what the data would look 16 like, and if we could have any sort of impact on our 17 clients. 18 So, we decided to start with a very simple 19 message for the feasibility study. We decided to see 20 if we could improve whole wheat bread consumption. 21 We had Catalina target all EBT users who 22 ate bread. We weren't trying to influence tortilla 101 1 eaters, for example, just white bread users, to go to 2 whole wheat bread, and whole wheat users to increase 3 consumption. 4 We had a very short intervention, what I 5 call a "light-weight" intervention. The messages 6 that were printed by Catalina, were very targeted. 7 I've got samples of them. This all happened at the 8 time when Catalina went to color, had these color 9 messages. 10 If you did not eat whole wheat bread, it 11 specifically talked about the benefits of whole 12 wheat bread for your family, specifically. We 13 reinforced this with our FSNE, but, keep in mind 14 that we may not -- all of these users may not have 15 been FSNE participants or SNAP participants. 16 I was delighted with the quality of the 17 data, and, specifically, the differences that we 18 were able to show with this very non-intensive 19 intervention. 20 We were able to look at data by users, by 21 ounces of whole wheat bread, and by the loaves of 22 whole wheat bread. 102 1 We also looked at displacement. We 2 looked at white bread, so we had the ability of 3 looking at over 500 USP food codes, and we had 4 identified those that met our criteria. 5 So we tracked all the food codes and then 6 all of the users. We ended up having 10,000 bread 7 users who used SNAP ed, EBT. 8 MR. CARLSON: Have you written up the 9 results of this? Are they available somewhere? 10 DR. TOWNSEND: I'm doing that right now. 11 MR. CARLSON: Could you be done by the 12 end of the afternoon? 13 (Laughter.) 14 DR. TOWNSEND: I have some results with 15 me. 16 MR. CARLSON: We'll be interested. 17 DR. TOWNSEND: There's one other thing I 18 will comment on. I have my contract here, with 19 Catalina, and I have my colleagues here, and I could 20 sort of share that with you, because it struck me, 21 when I was looking at our SNAP ed costs, versus our 22 cost for targeted messages, it seemed like a very 103 1 good use of the dollars. 2 So we purchased 130,000 targeted messages 3 for point-of-purchase delivery to the targeted food 4 stamp recipients. That was at ten cents a message, 5 so we paid Catalina -- this was a couple of years ago 6 -- $15,000 for the feasibility study. 7 MR. CARLSON: Thank you, Marilyn. 8 DR. PALMER: I think we could really 9 borrow from a lot of your techniques. I think it's 10 a really good idea. 11 I think that one of the reasons it was so 12 successful -- and I would like everybody to keep 13 this in mind -- is, I would suspect that people 14 largely replaced this product for that; they were 15 largely replacing whole wheat bread for white bread, 16 and I think that if we think that way in terms of 17 fruits and vegetables, again, in terms of, instead of 18 this snack, eat this snack, instead of this breakfast 19 have this one, I suspect we'll also be very 20 successful, especially if we use the kind of 21 techniques that you all used. 22 DR. TOWNSEND: I just presented this to 104 1 share the feasibility of the concept. I'm not 2 suggesting that we do whole wheat bread. 3 MR. DRUHAN: I just had a question. I 4 think this is something for the next couple of 5 panels to grapple with. 6 In Pennsylvania, we wanted to do a study 7 of people eating more vegetables, and we a lot of 8 trouble getting approval for it, initially, around 9 the preservation of personal information. It's one 10 of the problems of getting people to cooperate, is 11 being able to collect that kind of information on 12 individuals. 13 DR. TOWNSEND: We went through our IRB at 14 the University of California Davis campus, and didn't 15 have any problem with the ethical issue. We never 16 received any names or IDs. 17 Where we needed permission, since this is 18 proprietary data, was from our supermarket chains. 19 We ended up working with a chain called Rayleigh's 20 Nob Hill in California, and they were very, very 21 supportive. 22 We worked with their registered 105 1 dietician. 2 MS. GLASS: Could I make a comment on 3 privacy? I know we've done a lot of work on this, 4 but there's new technology that you can do with 5 prescription targeting, so you can imagine there's a 6 lot of compliance things we've done with that. 7 We do have a system now where we take any 8 card and it creates a unique ID within our system, so 9 there's not a firewall. It no longer leads to 10 consumer-specific information; it's always behavior. 11 It all gets created in a unique PIN code 12 in our data. 13 MR. CARLSON: I really hesitate to cut 14 off discussion, but we've got a long queue of folks 15 back there, and lunch is approaching quickly. 16 This is an area that needs to be pursued 17 or can be pursued more, and we'll all be around. 18 Next. 19 DR. HERSEY: Jim Hersey, RTI. I have a 20 question for the panelists. 21 We want a sustainable program, and one of 22 the things I've seen is, you will charge a surcharge 106 1 to people who don't participate in the program and a 2 reduction in cost for people who do participate. Is 3 it feasible for this study to conceive of something 4 like a five-percent surcharge for heavily-processed 5 vegetables and beverages, to then increase the 6 reduction in cost on fruits and vegetables? 7 DR. DEVANEY: My point is similar. Could 8 I add on? Barbara Devaney from Mathematica Policy 9 Research. 10 It seems to me that what has been talked 11 about, is, if there's a SNAP program that has spent 12 a certain amount on the targeted food, then it's a 13 bonus. 14 Alternatively, you could think of just 15 giving an additional benefit. You could think of 16 doing, for every SNAP recipient, giving them an 17 additional $10 worth of benefits, that would be for 18 the targeted food, which is different. 19 A third way is to say, the more you spend 20 on the targeted food, the greater your bonus is. I 21 wasn't quite sure. 22 The fourth way is what Jim was just 107 1 saying. Maybe you have a penalty, if you don't buy 2 healthful food. 3 Has it been decided, what the actual 4 design is, or any thoughts about that? 5 MR. CARLSON: I can answer that. No. 6 (Laughter.) 7 MR. CARLSON: Although I do think it is 8 true that the Department, as both Jessica and Katie 9 mentioned earlier, is much more interested in 10 providing a positive incentive to make a good 11 choice, as opposed to a sanction or a fine or a 12 negative incentive for a less good choice. 13 But the basic structure of the kind of 14 options that you laid out, Barbara, whether the 15 incentive is linked to a specific behavior or it's 16 just tied to a specific use, it's just extra income, 17 in effect; it's one of those design choices that 18 we're looking for advice and counsel on, and it's not 19 been decided and it's really critical where we go 20 forward. 21 DR. PALMER: I think it's important that 22 we not get into punitive measures, Jim, if for no 108 1 other reason -- this is a crazy thing I'm about to 2 say, but not everyone is overweight, and when you're 3 living in a household, again, with three teenagers 4 and you need to get calories into them because they 5 are losing weight, otherwise -- and this actually 6 happens, especially among teenage males sometimes -- 7 I would hate to see more money taken away, whey they 8 are already having a hard time keeping weight on 9 them. 10 DR. PIVONKA: I would have advocated for, 11 once they have a threshold of fruit and vegetable 12 purchases, they get the bonus, but I wouldn't be 13 opposed to the idea of, let's say, they're not using 14 their dollars properly, they're not using many of 15 their dollars on fruits and vegetables, and then we 16 target-market to them to try to increase what they're 17 doing. 18 I'd be comfortable, I think, with that. 19 In theory, at least, the ones who are saying they 20 spent $10 of their $50 on fruits and vegetables, are 21 already eating them. 22 On the other hand, we could work to try 109 1 to get those who aren't eating any, to get up a 2 little bit higher, so I would be okay either way. 3 DR. PEREZ-ESCAMILLA: Realistically, 4 taxing unhealthy foods, the way we tax cigarettes, 5 would be so controversial that I don't think that 6 would be doable right now. 7 MS. FOERSTER: Susan Foerster, with the 8 California Department of Public Health. We've been 9 blessed with legislation along this line, and so 10 we've been looking a little bit at how to do things 11 like this. 12 The last exchange was talking about the 13 money. For us, when we look at the legislation that 14 says four to seven counties -- they're big counties 15 in California, eight of them, but for just four 16 counties, it was about $10 million in food money, per 17 year. 18 That was before the fruit and vegetable 19 prices started going up, or food prices started 20 going up as they were. 21 So, I think that, whatever the design is, 22 we need to somehow take advantage of possible 110 1 natural market opportunities. 2 When we had the idea for this project a 3 number of years ago, our thought was that retailers 4 typically do promotions of certain types, whether 5 it's promoting a brand of salad mix or whatever it 6 happens to be, or if it's a seasonal kind of thing, 7 where they're pushing the apricots or pushing 8 whatever is available, if they could perhaps 9 preferentially offer those benefits to food stamp 10 customers. 11 The problem is, that's not allowed by 12 law, for them to do it. They can't charge a 13 different price, either higher or lower, to the food 14 stamp customer, so that really ties their hands for 15 even doing good corporate citizen work with food 16 stamp customers and taking advantage of the type of 17 market opportunities that are available to other 18 customers. 19 I think -- I wanted to ask Kathy Roork, 20 if they had really looked at more marketplace- 21 driven approaches, because I think there are a lot 22 of approaches that are buy-one-get-one-free, loyalty 111 1 cards, other types of things that operate in the 2 commercial marketplace, but which the food stamp laws 3 and regulations, prohibit. 4 Even in the execution of this kind of 5 study, I think it's not even a subject issue, 6 necessarily, that's going to limit this, but all of 7 the waivers and clearances and so forth, that would 8 have to be navigated with the food stamp program 9 itself, in terms of use of the data, which are owned 10 by the stores, and which, it sounds like, will be 11 very difficult to get full access to those data, to 12 evaluate those kinds of things. 13 If one looks in the market basket of food 14 stamp customers, the first part of the month, they 15 use their food stamps, they're gone by the middle of 16 the month, so, do you know what they are purchasing 17 over the month? 18 How do you tack those individuals, so you 19 can answer the question that we've been talking 20 about? So I think there are some other issues 21 related to the successful prosecution of something 22 for which there's $20 million. 112 1 Perhaps there needs to be a hybrid 2 approach that allows for basically the real world. 3 There hasn't been much discussion about the food 4 package, which is introducing fruits and vegetables, 5 period, fresh fruits and vegetables, period, and for 6 which stores are having to go out, state programs 7 are having to work together. 8 We still call them "food stamps" in 9 California, so I hope I don't get charged when I 10 use the term, but we are looking at a coordinated 11 approach to the WIC rollout, using the exiting 12 retail work that we used for FSNE or SNAP ed, along 13 with what WIC does with their vendors. 14 They have a regulatory role. They cannot 15 do a promotion, the marketing, the education, with 16 the retailers, that we can do. 17 So there's an opportunity to marry the 18 assets, if you will, of the different programs for a 19 more cohesive communication approach of the type 20 we've been doing, multi-level, multi-component, 21 where it hits the kids, it hits the moms. 22 You can do that if you marry the program 113 1 operations, and so I think that some of those 2 aspects, if we feel an urgency about getting the 3 program going, because of the public health reasons, 4 the chronic diseases and all of those other things, 5 we're going to have to look at removing barriers to 6 implementing the project. 7 Those kind of barriers can add one to two 8 years of time, if they're going to go the usual 9 course. 10 I guess that the last thing I wanted to 11 say, is, I think there are different approaches that 12 it can be delivered through -- and this is Kathy's 13 question -- it can be delivered through the EBT 14 system, or you have to change the EBT system, or 15 could it be delivered through the retailer, who does 16 this all the time? 17 Could it be where the food stamp program, 18 the SNAP program, reimburses retailers who use the 19 Catalina system or other kind of systems, who deliver 20 benefits? Maybe if they are on special, can they be 21 paid separately and have it put on the EBT card, per 22 se, for simplicity, at this point in time, or to 114 1 reduce the cost to the Government and to more 2 mainstream. 3 After all food stamp participants are in 4 a program for eight months. That's kind of an 5 average. I think we have to think and be very broad 6 and protective about what we're trying to do, but, 7 long-range, can the food stamp program do this 8 strategically, to do some fundamental changes in how 9 the marketplace works? 10 DR. PALMER: I can't answer anything else 11 you said, but with regard to retail relationships, 12 proprietary data, compliance with what we want to do, 13 using their system, we've been trying to do that in 14 New Jersey, and what we find, is, that we're able to 15 work with the corporate, so any corporate stores, any 16 company stores, will work with us, no problem. 17 But a lot of stores are franchised, and 18 it's just too difficult for franchised stores to do 19 this. 20 MR. CARLSON: I would just point out that 21 one of the challenges in thinking about the design of 22 this pilot, is to recognize the diversity of the 115 1 authorized retailer population in the SNAP program. 2 We have about 170,000 stores authorized 3 to accept food stamps around the country. Twenty- 4 percent of them are supermarkets; a smaller number 5 or percentage are grocery stores. A large majority 6 are small mom-and-pop convenience stores, farmers 7 markets. 8 The whole list of kinds of stores, goes 9 on for a full page. So, thinking about how do you 10 roll out a project, well, rolling out a project in 11 one segment of that population, has the risk of 12 appearing as if we're trying to promote that segment 13 of the retail population, and, on the other hand, as 14 I said earlier, 80 percent of our benefits are 15 redeemed through the large supermarkets and grocery 16 stores. 17 So it's not clear how much of a real 18 promotion that would be, but it is one of the 19 challenges. Kathy, Sue directed a couple of 20 questions right at you. 21 MS. ROORK: Okay. The first question 22 about the market-based approaches, we did hear from 116 1 a lot of retailers that, like you said, they do 2 promotions, every day, every week, so building this 3 into something they already do, is preferable. 4 We also heard about the fact that a 5 barrier to do this right now, without a pilot, would 6 be that they can't treat food stamp participants 7 differently than everybody else. 8 But with this pilot, I understand there's 9 the opportunity to test this out. I don't know, in 10 terms of the future, what would need to change, if it 11 was permanent. 12 And then about the various options, there 13 are a lot of different options related to -- you 14 couldn't build it through paper and outside of the 15 EBT system. Someone did bring up -- a few people 16 brought up the Catalina model, and the only concern 17 there, was only the large supermarkets have a 18 printer, which would, as I call it, spew out the 19 coupons. 20 But there are a lot of different options 21 and different disadvantages and advantages for each 22 of the options. 117 1 MR. CARLSON: Yes, ma'am? 2 MS. KIERNAN: Joyce Kiernan, representing 3 the Advisory Committee on Public Policy of the 4 Society of Nutrition Education of the Food Security 5 Subcommittee. 6 I have some random comments. One of the 7 things that we talked about today, is fresh or 8 frozen. We didn't mentioned dried at all. 9 In looking at the way food can be 10 measured, we didn't discuss at all, the servings of 11 fruits and vegetables, as a way to describe what 12 that product might be, how that product might be 13 defined. 14 There is the possibility of using a tier 15 of benefits, so that a higher benefit could be 16 given, or a percentage of the total purchase of that 17 person would qualify for a bonus. 18 Just like I used Google for directions to 19 get here yesterday, I think a recipe is a way that we 20 can help people to learn how to get there, the 21 techniques, as well as the use of fruits and 22 vegetables. 118 1 Right now, we're in economic 2 circumstances where everyone is eager to make the 3 most of their money, and I'd like to hope that we're 4 all out of this economic push before this gets 5 implemented, but we're all going to hold our breath 6 on that one. 7 We're all accustomed to marketing, we're 8 used to people cutting coupons out of the newspapers 9 or the kiosks at the store. We're used to getting 10 those coupons, so I don't think that's a barrier. 11 I think that could be a benefit. It is a 12 barrier, if you don't have those. 13 I think it's important to look at the 14 cost/benefit analysis on this, as a way to making it 15 sustainable, long-term. 16 We didn't talk about seasonal foods. 17 We've got all the fresh farmers markets, but we also 18 need to remember that they really -- 19 MR. CARLSON: Could you send me a 20 reminder in an e-mail in a couple of months? 21 (Laughter.) 22 MR. CARLSON: Thank you. Any reactions? 119 1 Random reactions to random comments? 2 MR. MOONEY: Random reactions to using 3 servings as a measure? We have pretty much move 4 away from the five-a-day, which became seven-a-day 5 and nine-a-day, and now it's simply eat more, 6 because no one can really tell what a serving is. 7 That's my random comment on that random 8 observation. 9 MR. DRUHAN: I'll comment on two things: 10 Low-income consumers often don't go to supermarkets 11 on promotions. 12 MR. MOONEY: Very often, the stores that 13 are patronized by low-income, are not equipped to 14 handle food coupons and don't even accept them. 15 DR. PALMER: Not only that, Kathy said 16 this before and I'll say it again, paper can be 17 sold; paper is sold, continually, so, within the 18 audience, there might be enough money for food, but 19 there's no money for other things. 20 DR. DISOGRA: Lorelei Disogra, United 21 Fresh Produce Association here in Washington. I'd 22 like to highlight some and add on to what Tim said 120 1 about the pilot studies that were done. 2 I think they vary a lot. There were 3 studies done on adding a voucher to WIC in a 4 controlled environment. We've done it in New York. 5 This is starting in California, just like Ellen said. 6 In 2001, one was in south central L.A. 7 That study has been written up and published twice, 8 most recently, last year, in the American Journal of 9 Public Health. 10 That study looked at -- they did it in 11 four supermarkets in a very controlled area of south 12 central L.A. They got vouchers, $10 a week, $40 a 13 month. 14 I believe the intervention was a year, 15 and that will give you a sense of followup, as well. 16 Ten dollars a week is certainly a lot of money, but 17 that was for the whole family. 18 They also measured behavior change. They 19 saw this as additive. So the vouchers, they can 20 substitute. They added more fruits and vegetables to 21 what they were buying. 22 The other study was in a very rural area 121 1 of California, Calaveras County, up in the 2 foothills, which had a whole variety of different 3 retail vendors, only one supermarket, about ten 4 different very small markets, and many mom-and- 5 pop's. 6 So, when you think about it, in whatever 7 area and how this is going to work, they made those 8 vouchers work in all those different retail vendors, 9 and that was $5 a month. 10 Tim, yours was $5, for kids only? So, 11 what we saw in this study in California, very 12 similarly to what Tim did in New York, is that the 13 moms really valued these vouchers. They were 14 redeeming 91 percent. 15 I heard a lot of discussion this morning 16 about, with low-income families, what would they do 17 with these vouchers? That's not at all what the 18 experience was with what was found in California or 19 New York. 20 They knew exactly what to do with these 21 vouchers. They were desperate to buy more fruits 22 and vegetables. In the two studies in California, 122 1 it didn't matter, and, across the board, the knew 2 what to do with these vouchers. 3 We were able to analyze, with the help of 4 UCLA, what was purchased in the two California 5 studies. They purchased a wide variety of fruits 6 and vegetables -- a wide variety, and in asking moms 7 that were part of these pilots, you know, how could 8 they explain what they purchased and what was 9 important to them, the moms would say, the WIC moms 10 would say, we were finally able to buy fresh fruits 11 to give to our kids as snacks, again, an opportunity 12 that was mentioned, and green vegetables for dinner. 13 They didn't get any special nutritional 14 education in either one of these projects. They 15 were already in WIC and got the regular WIC 16 nutritional education, 17 So, you know, I think we really need to 18 look at the experience from WIC and these three 19 pilots, learn from that and apply that for an 20 example, which is, I believe, not hard to do at all. 21 And we're talking about a research 22 project, and I don't think the Calaveras study has 123 1 been published, but certainly folks in USDA and WIC 2 have those results. Thanks. 3 DR. PEREZ-ESCAMILLA: Essentially what 4 you're saying, is that, based on the WIC experience, 5 this is not such an irrational idea. 6 DR. DISOGRA: Right. Going back to 1991, 7 one study in California, is still going on. They're 8 still giving those moms, the $5 vouchers, every 9 single month, $5. Not a lot of money in 2001 or a 10 lot of money now, but we've done it for three years 11 with kids, for several months, using state money for 12 that. 13 I think it's something that there's a lot 14 of lessons that can be learned and can be applied to 15 this research project. Thanks. 16 AUDIENCE PARTICIPANT: You said in your 17 study that there were no substitution effects; they 18 didn't really see any? Are you telling us that if 19 you give them $5, let's say, for vouchers, that the 20 total consumption of food went up by $5 and all of 21 that depends on the extra $5 of marginal increase? 22 DR. DISOGRA: They could only use the 124 1 vouchers for that. 2 AUDIENCE PARTICIPANT: They have $5 3 increase. 4 MR. CARLSON: This might be a good 5 conversation to have on the other side. 6 DR. DISOGRA: It's really set up as a 7 research design that has the best results. It was 8 done at UCLA, in south central L.A. That is the one 9 you should really look at. 10 They measured behavior change, and, 11 again, going for six months after they stopped the 12 incentive of giving the WIC moms the vouchers. That 13 one was $10 a week for a family, $40 a month, in 14 total, and that was published. 15 MR. CARLSON: Thanks, Lorelei. Yes, sir? 16 MR. THOMAS: My name is Ben Thomas, the 17 Food Policy Coordinator of the New York City Mayor's 18 Office. 19 I have a comment, which I guess is sort 20 of a preamble to a question. The comment is, Steve, 21 you mentioned, when you think about how we design the 22 pilot, we should think about the cost-effectiveness 125 1 of doing this to scale. 2 As someone who is relatively new to the 3 food field, we should also think about how to solve 4 the problem. The problem is massive, as everybody 5 here knows. 6 What you might not see, those of you who 7 have been in it for a long time, is the public 8 awareness of the problem, is really, really growing, 9 because the problem is also kind of slow moving, and 10 I think we should, as kind of a community of people 11 working on the problem, think about how do we solve 12 the problem. 13 If it turns out to be really expensive, 14 we should be prepared to make that case, that this 15 is a serious problem, and when life expectancies 16 start to go down in this country, the investments 17 will come. 18 So, within that, I also think the other 19 point that was raised repeatedly by Tim, is about 20 sort of looking at this as a opportunity to break 21 down silos. If you talk about tacking on 22 nutritional education to this pilot, don't ignore 126 1 the nutritional education already going on at WIC, 2 and don't ignore the incentives that already exist 3 in your own program, through farmers markets and so 4 forth and so on. 5 Don't ignore the school meals program as 6 an opportunity to tack on additional education, 7 because, sitting at the municipal level, it looks 8 like there's all this stuff going on, and it's 9 really hard. There's sometimes a lot of resistance. 10 And don't ignore the independents, funded 11 not at all through the USDA or the Government. 12 Within that, the question is, I'm curious to know if 13 people on the panel, who talk about providing the 14 pilot, the incentive, with education, and you talk 15 about our problems with access -- 16 In New York City, we have a lot of 17 independent efforts to promote access to fruits and 18 vegetables, where there is access. We didn't talk 19 about, well, what about combining the incentive with 20 efforts, educational efforts and efforts to promote 21 access. 22 DR. PEREZ-ESCAMILLA: I terms of access, 127 1 I don't think this program can work, if the issue is 2 not addressed. There's no way the incentive can 3 work, if there is no access to the participants that 4 depend on it the most. 5 In our cities, which is where I have 6 worked, it's a major issue. I fully recognize that. 7 In the rural case, it is, as well, so 8 it's actually a very important issue that has to be 9 addressed, first and foremost. 10 MR. MOONEY: I would fully agree with 11 that. In reading through the GAO report, I noticed 12 that there are some minimal standards that are 13 required of SNAP retailers in the foods that they 14 carry. 15 In the WIC program, there are much more 16 rigorous standards that are required in order to 17 participate as a WIC retailer. One of them will be, 18 in New York -- perhaps not in every other state -- 19 you will have fresh vegetables and fruits in your 20 store. 21 If you do not, you will not be accepting 22 WIC program checks. It is that simple. 128 1 What we have heard from the retailers 2 associations, is that if you make that expectation 3 known to the retailers, they will make the foods 4 available. 5 In New York City, food stamps and WIC are 6 critical to the existence of that store. Lose food 7 stamps, lose WIC, sell your store. So, they are 8 willing to make these additional foods available, if 9 you have the expectation of them. 10 It is not just a demand; it is also the 11 expectation that you have of them, in order to make 12 these healthy choices available. 13 MR. CARLSON: I do feel compelled, as an 14 FNS employee, to say that there has been an historic 15 basic difference in the approach to WIC, that we take 16 to the retailer community, in part, driven by the 17 nature of the program. 18 WIC is a much more prescriptive program, 19 a limited set of food items is prescribed, largely 20 to meet the special needs of women and young 21 children. 22 Food stamps is to a supplement benefit to 129 1 purchase, in addition to household income. The 2 criteria to get into WIC, to be an authorized 3 retailer, is approached more as a privilege, and 4 with food stamps, it's by law, basically a right. 5 If you meet the standards set by our 6 regulations, you're enabled to participate. 7 That is largely, I think, designed to 8 maximize the freedom of consumer choice among low- 9 income folks that we serve. Everyone here has the 10 same ability to go to Whole Foods or 7-11, and 11 there's been a longstanding approach, but there's no 12 reason why low-income folks should be treated any 13 differently. 14 There's also a pretty extensive evidence 15 that people shop in a variety of places, for a 16 variety of reasons. Sometimes we go to Whole Foods, 17 sometimes we need to stop at 7-11 on the way home. 18 MR. CARLSON: Since I put the systems 19 panel on the spot earlier, let me put the following 20 evaluation panel on the spot, as well. We've heard a 21 lot of observations about the importance of 22 combining, reaching out to a variety of settings, 130 1 programs, and other actors in all of this; testing a 2 variety of levels of incentive in a variety of forms 3 and a variety of ways. 4 The evaluation challenges that poses, how 5 do you know which of all those things, actually made 6 a difference? Do you have to do all of that, or can 7 you do just one of those things? 8 The more you put in the pot, the harder 9 it is to tell what actually contributes to success. 10 I don't have the answer to that question. Those of 11 you who are coming back after lunch, be prepared. 12 Yes, sir? 13 MR. KLEIN: Mitch Klein, Krasdale Foods. 14 The panel mentioned couponing throughout the 15 discussion. 16 With couponing, we talk about the health 17 issues and the lack of fruits and vegetables and the 18 lack of consumption. When we talk about that, we do 19 see residence come to the forefront of how folks view 20 the lack of fruits and vegetables. 21 With that type of recipient in the inner 22 city, they happen to have the absolute lowest coupon 131 1 redemption rate. They are the hardest people to 2 track. 3 They have the poorest response. If you 4 talk about general types of couponing in the 5 industry, more than 50 percent of your inner city 6 retailers, are on will-pay status with CMS, NCA, and 7 so forth. 8 If you go the route of couponing, aren't 9 you cutting off the people who you probably want to 10 most bring into this system? 11 MR. MOONEY: Mitch, I think that's a 12 great thing for you to say this afternoon, again. 13 MR. KLEIN: Thank you, Tim. 14 (Laughter.) 15 MR. MANALO-LeCLAIR: George Manalo- 16 LeClair, the California Food Policy Advocate. 17 Along the way, I've learned the economic 18 choices that drive the design choice. I just wanted 19 to ask the question, which sort of relates to this 20 advocate's perspective on what he'd like to see in 21 the pilot. 22 With the $20 million, everybody's doing 132 1 the math, trying to figure out how that would work, 2 and I'm just wondering if anything prevents us from 3 utilizing the large amount of private foundation 4 giving in the area of prevention. 5 I think there's a lot more interest 6 outside of just this effort, that we could pull 7 together. I want to make sure we have additional 8 resources, because the design choices that have been 9 discussed, trouble me, as an advocate, in some ways. 10 For me, I would love to see the intent to 11 go see everybody, regardless of who you are and where 12 you shop. The more we talk about the household, the 13 more concerned about privacy and other aspects of 14 this, and then about -- I want to make sure that, at 15 least from our perspective, this wouldn't be directed 16 elsewhere. 17 Can we talk about pulling together, the 18 public/private partnerships? It's an area that 19 would help us, bring us together, so that we don't 20 have to make some of those choices. Thank you. 21 MR. CARLSON: I think that's an 22 intriguing idea that we talked very briefly about. 133 1 We'll have to have some work on this issue. 2 Part of the issue is that it needs more 3 conversation. 4 (Laughter.) 5 MR. CARLSON: The observation I was going 6 to make, is that I don't think it eliminates the 7 difficulty of the choices; it just gives us a little 8 more flexibility than it would have otherwise. Yes, 9 sir? 10 MR. BALMER: I thought I'd have go get up 11 and respond to Mitch. I don't disagree with you. 12 I'm Clayton Balmer, Catalina Marketing. 13 I think there is a broad approach to take 14 here. You take a different approach in different 15 marketplaces. 16 When you look at how we go to a 17 marketplace, it's an integrated approach. We don't 18 need the same approach across the marketplace. 19 What we've learned at Catalina, is, we 20 have shown, through this very question about the 21 Hispanic marketplace -- as matter of fact, I can 22 talk about this to a greater degree later, so I'm 134 1 not interrupting lunch here -- but we actually had 2 people who said that the Hispanic consumer does not 3 redeem offers, they do not redeem coupons, there's 4 no reason to speak to them through coupons. 5 We found, actually, through CHR and 6 another study that we've done, that Hispanic 7 consumer redeem coupons no differently than anybody 8 else. They may redeem them at a slightly lesser 9 rate, but if you speak to them, if you speak to them 10 bilingually, they redeem offers, just like anybody 11 else. 12 They actually appreciate the value, and 13 we have shown, through a very clustered group of 14 stores, that when you talk to Hispanic consumers -- 15 and I'm not suggesting that Hispanic consumers are 16 the only ones on the SNAP program, but you might 17 suggest that some Hispanic consumers, especially 18 with the basic mission I'm responsible for at 19 Catalina, which is a lot of regional and independent 20 retailers in the West, these are sometimes one-store 21 mom-and-pop retailers -- when you help them speak to 22 their consumer in a very specific way, you can be 135 1 very successful with their consumer, as well. 2 So my recommendation is that you don't 3 need to look at this as just a coupon. You may not 4 coupon in the inner city, because couponing may not 5 make sense to some of the inner city shoppers, based 6 on where they shop, if they shop a bodega, if they 7 shop a corner market. 8 I think you made the great point, Rafael, 9 about access. You've got to have the right 10 products, and you have to have the right mix of 11 communication for them. Thank you. 12 MR. CARLSON: Let me offer the panel one 13 last opportunity for the last word. 14 (No response.) 15 MR. CARLSON: It's a limited opportunity. 16 MR. DRUHAN: I'll just say one thing, 17 that we tend to have a password for nutrition 18 education. In some states, they're lucky enough to 19 have governments that put direct dollars into the 20 programs and draw down the match that way. 21 Other states don't do anything and some 22 states do it to force nutrition education in a 136 1 particular direction. We should not talk to the 2 Feds about getting money; we ought to talk to our 3 states about matching for employee programs. 4 DR. PEREZ-ESCAMILLA: I just want to say 5 that for many years, I have been a strong advocate 6 for the change in the name of the food stamp 7 program. I've always advocated for a name that has 8 something to do with public health. 9 I'm extremely happy that we're talking 10 about fruits and vegetables, and now the name of the 11 program is SNAP. 12 MR. CARLSON: Thank you very much. Can I 13 just say that I think this panel has set a very high 14 standard for the rest of the day. Thank you very 15 much, and I thank all of you for your endurance so 16 far. Don't give up, we're only halfway through. 17 We are breaking promptly, as promised, at 18 11:45. The lunch options have not improved since I 19 first spoke. 20 (Laughter.) 21 MR. CARLSON: But we will plan on 22 starting up again promptly at quarter to 1:00. 137 1 We'll see you then. Thank you all very much. 2 (Whereupon, at 11:45, the symposium was 3 recessed for luncheon, to be reconvened this same 4 day at 12:45 p.m.) 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 138 1 AFTERNOON SESSION 2 (12:45 p.m.) 3 MR. CARLSON: If everyone could find your 4 seats again, we're going to do our best to reward 5 promptness. 6 (Pause.) 7 MR. CARLSON: I have a couple of small 8 housekeeping details as we get started. Those of 9 you who asked for assistance to get a cab or taxi 10 back to your destination after the close of business 11 today, whatever time the close of business is for 12 you, that has been arranged. 13 All you have to do, is go out to the 14 front of the building, that way, and there's a 15 little cutout there where the bus pulls up, and 16 that's where cabs are, so they will pull right up in 17 front. It's on the Plaza level. Take the P-Level 18 elevator. 19 Those of you who still need a cab, but 20 haven't done so, please sign up or see Kim at the 21 back, and she'll make sure to arrange that for you, 22 and get you where you need to go. 139 1 I'll also mention that we do still have 2 an interpreter, and if any of you in the room needs 3 interpretive services, it would be best to shift to 4 this side. They will be here to provide that 5 service, if needed. 6 With that, show of hands, does anyone 7 need assistance? 8 (No response.) 9 MR. CARLSON: Otherwise, she'll sit down 10 and enjoy the conversation with the rest of us. 11 (No response.) 12 MR. CARLSON: Maybe she'll practice. 13 Thank you very much; we appreciate it. 14 We're going to dive right back in. As 15 the topic of the first panel this afternoon, is on 16 the evaluation requirements and implications of what 17 we talked about this morning, we have an equally 18 distinguished panel of people to join us and share 19 their views, thoughts, and opinions with us. 20 For a change of pace, I'm going to start 21 at this end of the table and work my way up to the 22 other side. We'll start with Tom Baranowski, 140 1 Professor of Pediatrics at the Children's Nutrition 2 Research Center associated with the Baylor College 3 of Medicine, and is also affiliated with the 4 Agricultural Research Service at the USDA, so he's 5 almost one of us. 6 John Kirlin, is a economist, the 7 Assistant Deputy Director -- we changed your name 8 already, to the Special Nutrition Assistance 9 Program. 10 DR. KIRLIN: I wrote that. 11 (Laughter.) 12 MR. CARLSON: John works at the Economic 13 Research Service, also at the USDA. 14 Next we have Barbara Devaney, who is the 15 Senior Vice President and Managing Director of the 16 Human Services Research Component of the Research 17 Division of Mathematica Policy Research. 18 Next to her is Karen Glanz, who is the 19 Charles Howard Candler Professor in the Behavioral 20 Science and Health Education at the Rollins School 21 of Public Health associated with the Emory 22 University Prevention Research Center. 141 1 We've heard from Marilyn Townsend earlier 2 this morning, as part of the conversation. Just as a 3 reminder, she's with the Department of Nutrition at 4 the University of California at Davis, and is heavily 5 involved with the Cooperative Extension Service. 6 Next, we have James Hersey, the Sr. 7 Research Psychologist for the Research Triangle 8 Institute International; RTI, to those of us who 9 love acronyms. 10 Susan Foerster is from the Cancer 11 Prevention and Nutrition Section of the Network for 12 Health California, of the California Department of 13 Public Health. She also has knowledge -- I'll never 14 get the acronym right -- the Association of State 15 Nutrition Network Administrators and one of the 16 delivery mechanisms for SNAP. 17 Next is Gus Schumacher. He is a 18 consultant representing the Kellog Foundation, but 19 we in the USDA, also claim him as one of ours, in 20 his now former life. 21 (Laughter.) 22 MR. CARLSON: I was going to ask who you 142 1 were channeling today. 2 Finally, we have Miguel Pereira, who is 3 with Precima, Incorporated/Loyalty One. 4 We're going to follow the same format as 5 we did this morning. We have a series of prepared 6 questions to offer the panel. 7 We will address each of them in turn, and 8 we'll hope for some lively conversation among panel 9 members, as well, and there will be time at the end 10 of the session, for any questions or thoughts you 11 want to raise, based on what you have heard. 12 So, with that, let's just dive right in. 13 We heard several times this morning, that one of our 14 implicit goals and interests with these pilots, is to 15 be able to discern whether or not providing 16 incentives, don't just get used for their purpose of 17 fruit and vegetable purchases, but to do they 18 actually increase fruit and vegetable purchases 19 among the low-income SNAP participants that we're 20 targeting? 21 This is an issue that is described as 22 substitution versus supplementation, or are we 143 1 adding something to the overall mix or just 2 replacing. The question for the group, to start 3 with, is -- well, where do we find the data or 4 information to enable us to answer that question? 5 It would be great, we would think, at 6 least greater in the financial sense, if we could 7 find readily available data, as apart from purchased 8 data, that are already in existence or could be 9 readily acquired. It's less great, but perhaps 10 necessary to actually have to collect information, 11 which is a difficult and somewhat more expensive 12 proposition. 13 There may be other ways that we haven't 14 thought about. I hope that some of you have, but we 15 appreciate your views on, one, the importance of 16 being able to measure supplementation, as opposed 17 to substitution. Is it as important as we think it 18 is, to be able to measure whether households simply 19 substitute or do they increase fruits and 20 vegetables, and, if it is important, how do we go 21 about doing that? 22 DR. BARANOWSKI: I think this is an 144 1 enormous opportunity to really contribute to our 2 understanding of how the programs work or don't 3 work. 4 And if you miss the opportunity -- 5 families often don't eat by themselves. There's 6 other people in the home. How can we be sure that 7 they're not SNAP recipients? 8 There are multiple stores from which they 9 may be purchasing, so monitoring purchases at only 10 one store, may be inadequate. 11 There are food coops where people's 12 resources are just measuring purchasing. We can't 13 figure out where that stuff goes. 14 I think that measurement assumption is 15 absolutely critical. I would argue that food 16 preference questionnaires are absolutely 17 inappropriate and indefensible in this. 18 MR. CARLSON: That's why we invited you. 19 DR. BARANOWSKI: The National Cancer 20 Institute is developing one that will be available to 21 educators next year, and we can go multiple times to 22 their website and have very sophisticated 24-hour 145 1 recalls. 2 Whether it's valid or not, has yet to be 3 shown. 4 DR. KIRLIN: I certainly agree with the 5 Household Survey. I'm not convinced that it has to 6 be consumption, but we've got to go to the 7 households, because of the ability for people to buy 8 fruits and vegetables from non-EBT locations, with 9 non-EBT cash. 10 Once you're in contact, on the question 11 about expenditures, to get information on that 12 subject, I understand that's very difficult. 13 You also have to talk about the 14 evaluation issues as tools of research design. 15 Somebody's not going to be happy with the results of 16 the evaluation, no matter what that result is. 17 They will point to what they perceive as 18 flaws in the research design, so this whole issue 19 about substitution and product displacement, the 20 evaluation, really has to figure out some way to 21 deal with it. 22 MR. CARLSON: You're a much more ordered 146 1 group than this morning. 2 (Laughter.) 3 DR. DEVANEY: I have a couple of 4 thoughts: First, that's one of the reasons why I 5 asked the question about substitution. I think, 6 actually, before it's ever finalized, we really ought 7 to do pretty careful thinking through of what 8 substitution could be. 9 I also agree with Tom, that it's 10 important to measure consumption. I think of this 11 as the conceptual framework. The very first thing 12 you would expect to see from the pilot, is purchases 13 or expenditures. 14 And they want to see differences in 15 intake. If you only look at consumption or intake, 16 you never know, really, how much of it was because 17 of the substitution. 18 Unfortunately, it gets more and more 19 expensive to do expenditures and consumption, but, 20 in a pilot, if we really want to understand what's 21 going on, we have to do both. 22 DR. GLANZ: I agree with my colleagues, 147 1 but would just add a couple of other things. In 2 asking about food intake, I think the primary food 3 purchaser, is probably a good proxy for food 4 purchasing, et cetera, but as far as intake, they 5 might have very young children, so talking about 6 expense, if the whole household is SNAP-eligible, 7 different people are living there, it kind of goes 8 one step further. 9 It's important to get as much information 10 as we can, but it may be necessary to do some 11 sampling. One of the big things behind any issue, 12 is, what do you sample from? How big is it? How big 13 a power, et cetera? 14 It really struck me this morning, how 15 integrated these panels are. What are you 16 evaluating? What is the incentive? 17 There's a big difference, at least how 18 some pieces of it would be evaluated. I think, even 19 short of purchasing behavior, is the uptake and the 20 incentive. That has to be tracked along the way. 21 I think that's the first step. 22 MR. CARLSON: I haven't heard yet, any 148 1 disagreement, that measuring consumption, isn't a 2 part of this. As we go further on, I may, but if we 3 could add one element to that as well; is it 4 important to measure intake? If it's just as 5 important to measure consumption, is it important to 6 do it at the individual intake level? 7 The USDA used to do a household food 8 consumption survey, without worrying about who 9 consumed it. 10 DR. GLANZ: I'll point to Sharon's 11 comment this morning, about the number of people in 12 the household, is not going to be the same as the 13 size, age, and nutritional requirements of people in 14 the household. 15 Ideally, yes, there are fallbacks we can 16 make, and some compromises. It's probably 17 unrealistic, both in terms of gaining participation 18 in research, as well as actually getting everyone in 19 the whole household to allow you, on the online 20 system, especially for a low-income family, may not 21 be so easy. 22 DR. TOWNSEND: I have a couple of 149 1 comments. You referred to a national dataset. 2 Even considering that to evaluate this highlight, 3 the NHAINS collects data, the problem with that, is, 4 you don't know, within that diet, who used the EBT 5 card, so you have no idea what part of the diet. 6 We just finished a food cost study in 7 California, in conjunction with the University of 8 Washington. The clients that we studied, were staff 9 and participants, and approximately half of them were 10 on food stamps. 11 We collected food costs or food receipts, 12 for two weeks. If we're speaking of collecting food 13 costs, to get at food costs, I'm just telling you 14 that it was very difficult for food stamp recipients, 15 very difficult. 16 My next comment is another way we could 17 possibly get at food costs, is, the purpose of our 18 study was to validate a tool that would be easy to 19 use with large numbers of people, to assess food 20 costs. That happened to be the addition of a food 21 cost component to the food frequency questionnaires 22 by the Cancer Center in Washington. That should be 150 1 available. 2 DR. HERSEY: It's important to remember 3 that all substitution is not bad. Like the 4 randomized clinical trial I'm working on on the food 5 stamp population, where we're working on physical 6 activity, increase of vegetable consumption -- may be 7 a 30-percent increase of fruits, or 50 percent. 8 What's going down also the same 50 percent, is 9 snacking or junk snacks. 10 There may be useful displacement. We've 11 got to get at that. 12 Similarly, if I'm displacing the non- 13 food stamp portion of the food budget, am I eating 14 out in fast food restaurants? Those are some of the 15 issues we've got to get at. 16 MS. FOERSTER: On top of that, in the 17 food stamp program, all the states are working with 18 food retailers. To me, the central question is, are 19 we able to change the food stamp market basket, the 20 SNAP market basket? 21 There are commercial data systems that 22 could be used, such as IRI. We use something called 151 1 Fresh Look, that could be used to track all of the 2 purchases that are made when the SNAP EBT card is 3 used. 4 That's a worthy policy objective, to 5 understand if government dollars are being spent in 6 a different way. 7 In terms of going a little deeper and 8 looking at different stores, because we know people 9 shop at different stores, I would argue for kind of 10 a triangulation approach. 11 A couple of things: When I would argue 12 for a community-wide approach, where you could look 13 at all the stores in a community where you're doing 14 your doing your intervention, and see the general 15 shift of food stamp dollars, knowing people will use 16 them in different places, so I think there would be 17 some real issues with sharing data. 18 The companies have different price look- 19 up codes and that sort of thing, and that would have 20 to be dealt with, but that would argue for more of a 21 community approach. That's really the way that the 22 program, I think, is designed. 152 1 In terms of triangulation, I would hope we 2 could look at other data sources, PRSS, for example, 3 which all the states have. California has some 4 others. 5 We can ask whether a person has used food 6 stamps in the last 12 months, or have they used WIC, 7 so you can use the things where you can look at 8 different data to see if they're all going in the 9 same direction. 10 I think the issue of triangulation, is 11 going to be difficult. That's why you need to know, 12 what's the logic model that you created to figure out 13 how these effects work? 14 You do really need to be able to look 15 ahead at the things you were talking about in the 16 earlier panel, like what are the community factors 17 going on, and so forth, in order to draw the case, 18 and probably sophisticated statistical approaches to 19 tease out whatever you can on attribution. 20 MR. CARLSON: Let me just interject at 21 this point, that one of the things we'd like to be 22 able to do, that we've asked ourselves or convinced 153 1 ourselves we'd like to do, but haven't quite figured 2 out how, is, how do we rely on existing datasets? 3 One of the design features we haven't 4 decided on, is, what's scale the pilot is going to 5 be. Many of the national datasets that have been 6 discussed, NHAINS, for example, might work well, if 7 the pilot runs in a pretty large geographic area. 8 It it's run in a small community of a small part of 9 a large community, it's not clear to us yet, how you 10 would be able to link the data within those systems, 11 to the operation of the pilot. 12 Having said that, it's all sort of up in 13 the air at the moment. We don't know what the scale 14 of the pilot will be. 15 MS. FOERSTER: Whatever communities you 16 pick, I think -- the three studies I mentioned, were 17 all done in communities. You could look at all of 18 the outlets for all of them, and you could provide 19 the scanners for those who didn't have a 20 sophisticated enough system. 21 Then you could set up areas where 22 interventions were not being offered. In our 154 1 system, it wasn't felt that we could deliver an 2 incentive, except on a county basis. It has to be 3 countywide, so you'd have to have your comparison in 4 a different place. 5 MR. CARLSON: We're going to come back to 6 that and explore that a little more. Any other 7 thoughts on this? 8 MR. SCHUMACHER: Just briefly, one of the 9 things we have learned from our work at the Kellogg 10 Foundation, is, as long as we're making grants, we 11 would insist on ten percent. How much of the $20 12 million do you want to spend? 13 We insist that our grant be filled in on 14 two levels of evaluation: One, we provide that ten 15 percent of the grant must be spent on evaluation. 16 Each grantee has to select a evaluation substitute, 17 do an ongoing evaluation as part of the ongoing. 18 Then Kellogg itself has an independent 19 evaluation, so they would go cross-section and look 20 at the different grants and draw some lessons. I 21 think it's a question for you all to decide how much 22 you want to allocate. 155 1 The second issue: There's been a number 2 of lessons coming out of the WIC programs. 3 Originally, there was the 1992 WIC voucher program 4 we've worked on for 20 years. 5 More recently, for seniors, there's been 6 a lot of innovation by foundations on SNAP programs 7 that we mentioned in Hartford. There's been a lot 8 of incentive programs for EBT. 9 In early evaluations of those, some of 10 the lessons we picked up, show, surprisingly, in San 11 Diego, where you provided incentives for SNAP, what 12 they did, is, if you're not a SNAP participant, 13 they're, by percentage, one of the lowest users of 14 SNAP in the county. 15 San Diego's Promise Market, would give 16 you $20, just to come to the market to see if you 17 might be eligible. One of the things that somebody 18 mentioned, is word of mouth in urban centers. 19 What's most fascinating, was, they said, 20 SSI can sign up. There were lines of 200 to 300 21 people. We were stunned at that, absolutely 22 stunned. 156 1 It was embarrassing because of these long 2 lines of SSIs, some of them requiring food stamps, 3 some not, so there's cross sectional data, but 4 there's a lot of data coming out the market from WIC 5 and the seniors programs. 6 Finally, we discovered that 90 percent of 7 the seniors would use their vouchers; about 60 8 percent of the WIC. Then the early returns from 9 SNAP, just for the two years' anecdotal work, were 10 pretty modest, with very, very low utilization, even 11 with this voucher program. 12 So, it's going to be very interesting, as 13 is some of the work you've been doing, as well, 14 having to draw on some of those early lessons. 15 And what kind of expenditures do you want 16 to spend on evaluation? How do you want to 17 structure the evaluation? Do you do it 18 independently? 19 MR. PEREIRA: Just a quick comment. We 20 work with retailers across North America, many of 21 them grocers, and do a lot of work, taking their 22 customer-level data. This question about looking at 157 1 customers, versus an aggregate view, is very 2 important, especially if you're looking at a lot of 3 the discussion this morning about not just kind of 4 incenting people in a promotional way, or kind 5 building in price incentives that work in the short 6 term, but not the ultimate goal of it, and to be 7 able to track those individuals over time. 8 There was some discussion around Catalina 9 and how they do that. That's very important. 10 Secondly, we're talking about mainly 11 fruits and vegetables, but understanding that, the 12 individual behavior across the entire basket, is 13 going to tell you a lot about the shift. You just 14 don't want to increase fruits and vegetables. 15 A good indicator of that, is less 16 consumption in other areas. I think we were going 17 to comment on that, as well. 18 Really, at an individual level, is the 19 best way to get at that. 20 MR. CARLSON: That's a good point that 21 we're following up on as well. Jim, I think, said 22 it. 158 1 The nature of the substitution, has led 2 some of us to think that maybe we need to measure, 3 not just what people are purchasing, in terms of 4 food, but we need to know how else they're 5 allocating their money. 6 If the incentive for fruit and vegetable 7 purchases results in no overall change in food 8 expenditures, but results in more transportation or 9 medical care or a night out at the movies, would we 10 want to know that, as well? Yes, sir? 11 DR. BARANOWSKI: I'd have to argue that 12 you have go further. It was mentioned in the 13 session this morning, that there are different 14 interpretations about whether incentives work. 15 There's a whole controversy about do 16 incentives, in fact, induce a disincentive, when the 17 psychological rationale is that they have to pay me 18 to do this. 19 MR. CARLSON: You mean that incentives 20 result in snacks? 21 DR. BARANOWSKI: I think you need to go 22 out and measure fruits and vegetables, to identify 159 1 the issue of disincentives. If, over time, you find 2 you're not increasing fruit and vegetable 3 consumption, it is because people have learned its 4 value, that you're trying to promote? 5 MR. CARLSON: Presumably, that leads you 6 in the direction of a household survey, as well, in 7 order to answer those questions and probe their 8 motivation. 9 DR. BARANOWSKI: On a random-sample 10 basis. 11 DR. KIRLIN: We had an idea to give 12 everybody a $10 bonus, one group restricted to 13 fruits and vegetables, and the other can spend it 14 however they want to. 15 Anyway, there's a question of the method 16 having been validated. This may be an opportunity 17 to try to validate other measures and spend less 18 money. 19 (Laughter.) 20 DR. DEVANEY: I'd like to add something 21 about the data collection. I'm not sure on the 22 spending side or the cost side, what we'd get. 160 1 I know there are many other people in the 2 room that know this much better than I do. On the 3 intake and consumption, we know that they are very 4 expensive, even with, I guess, a web-based survey 5 that we're talking about. 6 If that were to be validated, even the 7 24-hour recall, gives very good data, but when you 8 have observations on nutrient intake, you're still 9 not so good on food intake. 10 Really what we're trying to get at here, 11 is whether there would be an increase in consumption 12 of fruits and vegetables. The messages there are 13 significantly less valuable and harder to use. 14 MR. CARLSON: The way the discussion so 15 far has generally been, has been for some form of 16 household survey -- not entirely so, but it's 17 generally been that way. 18 But let me pose the question the other 19 way: Aside from a household survey, is there 20 intrinsic value and is it achievable, doable, and 21 affordable to try to collect some broad measures of 22 aggregate sales in the area of the pilot? 161 1 DR. GLANZ: It comes back to design. If 2 you design it for counties or small areas that are 3 likely to be somewhat of an island of where people 4 obtain their food, that can be a great opportunity 5 in looking at the question on the way to study fruit 6 and vegetable purchases, versus fruit and vegetable 7 consumption. 8 My impression, at least from this 9 morning, is that we don't know that much. This 10 pilot could be a great opportunity for that. 11 It's one of the kind of things Tom is 12 talking about. There seems to be a lot of 13 opportunity here for self-studies or focus on 14 measurements or current validation of these things. 15 They can contribute not only for this 16 particular SNAP evaluation, but more broadly for 17 nutrition across the spectrum of whatever. 18 DR. HERSEY: I'm going to make the case 19 that I think you need to realize, if we're primarily 20 looking at fruits and vegetables, you need to think 21 about it from the respondent perspective. 22 The key things like carrots, green beans, 162 1 bananas, apples, people can remember those. What 2 you're not going to get from that, is total caloric 3 consumption, but if you were asked about key 4 indicators, they can tell us a lot. 5 Or there are some of the key methods of 6 food preparation. How often do you eat soup? 7 Then you use your storewide purchase data 8 to complement that by getting your aggregate sales 9 about everything. 10 You may have a greater ability to get the 11 kind of sample size you need to get relatively 12 modest effects. 13 MS. FOERSTER: The issue Tom was raising 14 about fruits and vegetables not being popular, or, if 15 you incent them, we're not seeing that at all. 16 What we see, is more what Augie was 17 talking about. People really want fruits and 18 vegetables. In our low-income audiences, it's about 19 42 percent of people that say that cost is their 20 biggest barrier, compared to maybe 12 to 15 percent 21 in the non-low-income population. 22 So, cost, for low-income people, 163 1 especially now, is much more of a barrier than any 2 kind of idea of they don't like them. And I don't 3 think we're very much different from other places. 4 The other question is, I don't think that 5 there's a lot of waste in low-income households, in 6 terms of buying fruit and not using it. 7 I am personally not familiar with the 8 research, but what we hear in our focus groups, is 9 that people are afraid of buying something that 10 either the family won't like and then wasting the 11 money, or that they don't know how to prepare and 12 would be wasting the money. 13 It looks to us like people are eating the 14 foods that they buy, or they don't buy it. So it 15 feels like to me, that the idea of waste, for 16 something like that, shopping, should be very 17 similar to consumption. 18 Again, this is a program that started 19 with retail. The retailers were interested in 20 selling more food and having more customers. 21 So I think that the idea of fruits and 22 vegetables being a fairly high-margin item, is also 164 1 something that's fairly appealing, so that we help 2 other parts of our economy. 3 Another issue of fruits and vegetables 4 and sales, is support for sustainable local 5 agriculture. That is very much a part of having a 6 reformed kind of food system. 7 So I think that there's a lot of assets 8 and there are other side benefits, not just finding 9 out how to do different sorts of tools and different 10 types of measures, but other side effects that we 11 ought to be looking at when this is designed. 12 I see some of these other economic 13 stimuli that the food stamp program has been shown 14 to provide. It's very important in low-income 15 communities. 16 MR. SCHUMACHER: Let me just follow up on 17 two points. Something you might think about a 18 little bit, in terms of the evaluation -- Lorie is 19 my counselor on this -- will you be doing or should 20 you build anything on price response? 21 Two things are occurring: We've been 22 watching basically four generations of fruit and 165 1 vegetable growers in the Northeast, and have been 2 watching what happens on price. 3 In the farm belt, you already have great 4 expansion of the fruit and vegetable programs in 5 schools. That's going to be about a billion 6 dollars. 7 DR. DISOGRA: It's $49 million. 8 MR. SCHUMACHER: Then you have the $50 9 million in other programs. In terms of the WIC, 10 that's going to be $432 million. 11 You begin to add it up, and we're looking 12 at a chunk of money. 13 Then you look at the supply response. 14 Can we build that into the evaluation? If you're 15 going to be putting in supply response in certain 16 areas, are the supermarkets or the smaller stores 17 going to be jacking up the price of, for example, 18 carrots? 19 You mentioned carrots. There's only two 20 real carrot growers in the country right now. They 21 grow 40,000 acres apiece in California, but what's 22 occurred, is, it used to cost $3,000 to ship a load 166 1 of carrots and tomatoes from the Valley, to Jessup, 2 Maryland or the Bronx, or to the Chelsea market; now 3 it costs $10,000 to run that semi-refrigerated 4 across the country. 5 Again, can there still be an evaluation 6 on some price responses? Are there systems for 7 piloting, that somehow keep a more regional system, 8 so that you don't have some price response? 9 You're going to have to allow us to get 10 quite clear in terms of the evaluation, but you're 11 going to have to look at a lot of federally-driven 12 demand for fruits and vegetables for lower-income 13 people in schools, in WIC, and perhaps is there 14 going to be the demand running up the prices, 15 because supply is going to be lagging? 16 MR. CARLSON: You might take some 17 comfort, I think, in the observation that the 18 markets for each of these three or four programs you 19 mentioned, are actually quite different. 20 For the fresh fruit and vegetable snack 21 program in schools, purchases are made by government 22 and institutions in the wholesale market, not in the 167 1 retail market. Seniors Farmers Market and the WIC 2 Farmers Market are really very small programs and 3 limited to purchasing in farmers markets. 4 What's new, is this one. Well, WIC is 5 sort of new in grocery stores, and this one. 6 So, it might be an issue there, but it 7 may not be quite as big an issue. 8 This calls to mind, two things to push 9 on: One is, it is fundamentally true that there is 10 just an enormous number of interesting and useful 11 questions to be pursued in these pilot projects. 12 We're just trying to hear some of them in all this. 13 I think we ought to be frank with each 14 other, that we can't do it all; it's just not 15 possible to do it all. 16 We're going to have to make some choices 17 about what is fundamentally important. I suspect 18 our first line of guidance on that, is going to be 19 the statutory language, where the focus really is on 20 increasing purchases of fruits and vegetables or 21 other healthful foods, in the context of addressing 22 the problem of obesity and overweight. 168 1 Having said that, our principal focus is 2 not to exclude the potential for looking at some of 3 these other effects, but we need to be careful and 4 not lose sight of the main goal here. 5 The other thing I wonder about, and a 6 question I posed to the panel, is this notion I 7 raised before we broke for lunch. Gus, your 8 observation of the variety of incentives in the 9 program, that the programs provided, reinforced this 10 in my mind. 11 There's a lot going on. There's a lot 12 more that could be going on, all directed in the 13 same direction, trying to increase or change fruit 14 and vegetable consumption and behavior. 15 Do we plot this set of pilots down, 16 isolated from all those others? Do we look for ways 17 to try to link up with those existing programs? Do 18 we try to link up? How do we know what value have we 19 added by that linkage, that is separate from 20 everything else? 21 I don't have yet, an answer to that. I 22 don't necessarily expect one. Maybe you have some 169 1 thoughts. 2 DR. KIRLIN: I'm going to pretend like 3 I'm in a debate and ask what you should have asked. 4 (Laughter.) 5 DR. KIRLIN: You could ask about the 6 reliability and necessity of aggregate data. I'm 7 really skeptical of that approach. 8 You're not necessarily saturating the 9 market with incentives; you're giving incentives to 10 a sample. You're giving a relatively small 11 incentive to a sample of participants. 12 When you aggregate that many people in 13 that market, it's not that big of a change. Along 14 with everything else that's going on, you think 15 about the leakage that's going to occur, trying to 16 measure aggregate sales of fruits and vegetables. 17 I just don't see it working. Measuring 18 the leakage out of your measure, is probably bigger 19 than the effect you're trying to find. I don't 20 think you look at those as a supply experiment. 21 Really, you're looking at it on the demand 22 side, on the consumer and what changes are being 170 1 made. 2 What was your question? 3 (Laughter.) 4 MR. CARLSON: The request for the thought 5 is, in the relative merits or importance of being 6 able to isolate the fact that this pilot, are we 7 better off it we integrate it with some other set of 8 interventions? Would this possible? If we 9 integrate it, how do we know what this pilot has 10 added? 11 DR. KIRLIN: I don't know what evaluation 12 you plan, but I could also -- in one of your other 13 questions, I think you've got to do this. 14 One of the reasons is that what's going 15 on in the communities, is often community-based. We 16 need to be able to control for that. 17 DR. DEVANEY: I would say that the first 18 thing we should ask of the pilot, is whether it can 19 be used to redirect or improve the food choices. I 20 think you need to set it up, so that you can really 21 test that. 22 You don't want to link it to WIC or other 171 1 kinds of incentives. I think, by the way, your 2 question actually leads very much into this other one 3 about community. 4 MR. CARLSON: They are related. 5 DR. BARANOWSKI: It's absolutely 6 critical. 7 In my area of the randomized clinical 8 trial, if you start reducing all of the 9 complications, statistically, you can play games 10 with trying to figure it out, but if you have an 11 opportunity here to do that, why would you want to 12 do anything else? 13 DR. GLANZ: I think one thing that sticks 14 in my mind -- I completely agree with my colleagues. 15 Having a clinical trial, is an opportunity to do 16 that, and it makes sense. There could be more than 17 one experiment in more than one type of community, 18 what have you. 19 But an RBT is not one thing. That, 20 then, is the next question. One can have a four- 21 armed trial with an increasing number of incentives, 22 which would be a much more efficient way to learn the 172 1 impact of different incentives, rather than just 2 trying out one and seeing it fail. 3 We can be trying other different things. 4 There was a lot of discussion this morning about 5 Catalina and its program. I'm not sure, within the 6 federal program, how acceptable the program is to 7 consumer's baseline, but it's a really intriguing 8 idea, and it would be an exciting thing to try. 9 It opens up everything for informed 10 consent, to incentives, I mean, there's a lot of 11 behavioral research that suggests that targeting of 12 some type, is more effective. This is somewhat new 13 territory. 14 DR. TOWNSEND: A couple of thoughts. 15 Although I agree with randomized clinical trials and 16 randomizing at the community level, that really 17 takes a lot of money. 18 We haven't discussed what the budget 19 would be. I don't know what part of the $20 million 20 goes to the intervention. 21 If there is a budget problem, I think, 22 for evaluation of a design of this type, we could 173 1 get good data from collecting. 2 Say we have two dozen communities 3 collecting baseline data for three months during the 4 time of year when we think it would be equivalent to 5 the intervention time. 6 For our intervention with Catalina, we 7 did just a one-month baseline data collection and a 8 one-month intervention. I know that's very 9 simplistic, but the concept of that, is infinitely 10 less expensive. 11 MR. CARLSON: Allow me to interject at 12 this point. One of the things that's unique -- and 13 I don't have the statutory language in front of me, 14 but all the rest of you do -- there is language in 15 here that puts more stringent expectations on the 16 nature of the evaluation than normally occurs in 17 federal legislation. 18 It's Section 4(a)(i)-something. It says, 19 basically, it shall use rigorous methodologies, 20 particularly random assignments or other methods that 21 are capable of obtaining valid information regarding 22 which activities are effective. 174 1 That language was subject to some debate 2 and discussion, but the sponsors of this language, 3 made it real clear that they wanted to be able to 4 answer this question. If it took a lot of money to 5 evaluate it, to follow that expectation, that was 6 okay to the folks who were writing this language. 7 There were others who were arguing that 8 more money ought to go to the intervention. We'll 9 settle for what we can learn from this language, but 10 I think we're reading this language and doing -- we 11 ought to start with the presumption that we're going 12 to do as rigorous an evaluation as possible, and then 13 back away from that, only as proves necessary. 14 DR. BARANOWSKI: I think you should have 15 the main study totally reverse the design, to see if 16 we can really nail this issue. However, there's a 17 lot of design issues that can be resolved. 18 One of the discussions should be, what 19 should the incentive be, $5 dollars, $10, or what? 20 There's a form of research whereby you 21 take a community and try maybe five different 22 communities and even different incentive programs, 175 1 to look at which one gives you the biggest bang for 2 the buck, however you decide you want to do it. 3 And that one is incorporated into the 4 main trial. All of the key issues in the research 5 design, need to go through some kind of rigorous 6 form of research. 7 That main study is -- 8 MR. SCHUMACHER: Is Tony still here from 9 New York? In some cases in New York, you're doing 10 $2, in some cases, it's $5 and in some cases, $10, 11 perhaps even $20. I think it's very interesting to 12 look at different levels of incentive as you're 13 doing your pilot. 14 One question I have for Steve, is, how 15 many pilots do you think you could do for $20 16 million? Should we be doing them as RFPs at the 17 million-dollar level? 18 In some of our foundation work, some of 19 the people make substantial grants and then do 20 rigorous evaluation. A question, maybe for 21 discussion later on: How many can you do? Will you 22 do them in three of the states, and will you then 176 1 offset that with some of the richer states? 2 There are all kinds of interesting design 3 questions. 4 MR. CARLSON: That's exactly right. 5 We're going to defer answering that question, 6 because, frankly, we don't know yet. We wanted to 7 have the benefit of this kind of conversation and 8 feedback. 9 MR. PEREIRA: I was going to make a 10 comment on site limitations. There seems to be some 11 sensitivity around that profile and targeting the 12 incentives. 13 Having done some of this work, we're 14 doing this, obviously, to understand how to price. 15 When we've tested it against a segment of the 16 population or the customer base, you get very 17 different price responses across each segment. 18 If you can do this and assume that SNAP 19 members are going to respond to prices in a sort of 20 aggregate fashion, is not the way to structure it, 21 but understand the sensitivities around that. The 22 ranges are across category. 177 1 The problem with looking at this, is, 2 it's not an issue of the product, per se, fruits and 3 vegetables; this is a people issue. You're trying to 4 get people to change behavior. They're not just 5 switching between buying things here, versus 6 somewhere else. 7 It's really a fundamental change. 8 Segmenting and understanding what's driving those 9 changes among different participants, I don't think 10 we're going to be able to answer that effectively. 11 DR. DEVANEY: I just wanted to interject 12 into the discussion of the design. Many people are 13 assuming that it has to be at the community level. I 14 don't think that's true at all. 15 I think you use your dollars a lot more 16 effectively, if you do it at the level of 17 individuals. Communities are very expensive; 18 there's a budgeting issue. 19 So I would really try to design it so 20 some things happen that you can look at individuals 21 within a community. 22 DR. BARANOWSKI: There are big problems 178 1 with this. The major problem is contamination. 2 People in the community talk to one another. 3 Once they start talking with one another, 4 there's going to be comparisons, like why does he get 5 that and I don't get that? 6 That's a huge problem. 7 DR. HERSEY: I think there's much appeal 8 in the individualized framework, but we need to 9 realize what we give up when we do that -- good 10 baseline data, people can have their own narratives. 11 But some of the questions I'm really 12 curious about, are if what I'm trying to do, has 13 increased consumption of fruits and vegetables. 14 Am I going to do that better by simply 15 subsidizing and adding ten percent, or would I do 16 better to give a ten-percent bonus card, or would I 17 do better with a coupon or promotion? 18 Some of those kinds of things, I really 19 want to know about, and are really going to require 20 implementation. It's not at a community level; it 21 may be at the store level, community level, or food 22 stamp office level, but I don't want to give up the 179 1 ability to capture some of that. 2 What's the best way to do this? 3 DR. GLANZ: Jim's comment actually plays 4 into something else that kind of came out. We were 5 talking about the scope and how much money people 6 would have to spend, how much on evaluation. 7 Two million dollars will not buy very 8 much. It will buy a moderate sized one offset. It 9 won't do anything to give you a definitive answer 10 about whether or which kinds of incentives work. 11 One question that we really haven't 12 talked about, is, do we expect the researchers, the 13 evaluators, to actually implement the incentives, as 14 well? It seems to me that before it becomes 15 systematic and institutionalized, whether in this 16 kind of program, somebody's got to implement it. 17 If something is at the study level, you 18 know, an up-front coupon for a purchase has value. 19 So, you know, there's a lot of questions 20 there. I think the way a lot of us who are NIH- 21 funded, we do the intervention and we do the 22 evaluation, as well, making sure there are 180 1 safeguards built in, so there aren't biases. 2 But that's a big project. It's a large 3 undertaking. 4 From the GAO report, I think it's clearly 5 an overriding need, not to have just a good solid 6 contract, but a grant to whoever is implementing it 7 could invite oversight from different groups. 8 There are a lot of different things they 9 want to clean up along the way. 10 MS. FOERSTER: If I could jump in, to me, 11 this is really a policy study. It feels to me like 12 you could define it as what do we want to know about 13 the ten percent, 20 percent incentive, versus how do 14 we want to change the food stamp program and how do 15 we want to change how that functions with the other 16 food assistance programs? 17 What we heard this morning, is that you 18 need multiple approaches. You're going to get more 19 impact, more synergy, if you look at these multiple 20 other assets that you have in your repertoire. 21 To me, yes, it's contamination, but it's 22 good contamination. You should be designing your 181 1 intervention, or the study, so that it takes full 2 advantage and treats child nutrition and so forth, 3 farmers markets, as part of the mix that you're 4 trying to get to a healthier food environment for 5 those people. 6 So you design those in the 7 considerations, up front, of those factors, and you 8 track them in the location that you choose. 9 In terms of community, partly, I'm 10 thinking, in my state, if I had to -- I would have a 11 very hard time finding a place that's free of 12 contamination. I think that's really true in much 13 of the country, because FSNE is really prevalent. 14 There is a lot of effort going on there, 15 obviously, with the child nutrition programs. 16 Finding a place that's pretty free of contamination, 17 where you could just look at just one or two things, 18 is going to be really, really difficult to find. 19 Then you have to be able to deliver the 20 benefits, so, it seems like a more naturalistic 21 design, than a randomized clinical controlled trial, 22 is really needed. 182 1 It's more of, maybe, a social sciences, 2 or more of an economic or more of a policy kind of 3 study, that is really using other kinds of methods, 4 because it is, after all, an entitlement program. 5 You can't really distinguish, at least according to 6 current rules, you can't give some people in the 7 program, a benefit that you can't give other people. 8 Like I said, in our situation, we would 9 have been able to give everybody in the county, the 10 same benefit, but we didn't think we could give one 11 person in the county, a ten percent, another a 20 12 and another a 30. We didn't think that would be 13 allowed. 14 The issue is finding comparisons 15 communities. 16 MR. CARLSON: Again, I'll exercise my 17 prerogative to interject with two observations: 18 One, the fact that the authorization for the pilot 19 has been written into the statute, we might have had 20 the authority all along, but this gives the 21 Department a little greater flexibility and ability 22 to think about options that we might not have 183 1 thought about, absent the specific directive to run 2 this pilot and to run it in this fashion. 3 I also want to pose to the panel, the 4 question that Sue is raising here. It's really hard 5 to avoid contamination, because, in any given 6 community, there's any variety of combination of 7 things going on. 8 Isn't that, in some sense, an argument 9 for, if it were feasible, what is the value of 10 randomizing at the individual level, would be an 11 important additional thing to do? 12 Within the community, they would have all 13 been exposed to the same set of factors, with the 14 exception that some have gotten disincentives and 15 others would have not. It's another way of dealing 16 with the problem. 17 John? 18 DR. KIRLIN: I think we're confusing two 19 kinds of contamination. One is the notion about 20 general information out there in the community. 21 That's always going to be there. 22 I could very well differ from one 184 1 community to another, which is why I'm not wild 2 about assignments by community. 3 The contamination I had referenced 4 earlier, was, you have a treatment group and a 5 comparison group, and you want to be able, if the 6 comparison group doesn't somehow get part of the 7 treatment effect, or vice versa, I think a 8 randomization within a community, having people 9 talking to one another, you've got to figure out how 10 to deal with that problem. 11 I think that problem is more amenable 12 than the problems that get introduced, if you were 13 never to assign by communities. 14 DR. BARANOWSKI: It's unlikely that you 15 would be able to do a whole study within one 16 community. You'll probably have to go to multiple 17 communities. 18 Once you go to multiple communities, they 19 need to be clustered, and there are statistical 20 aspects of clusters that violate normal distribution 21 and other kinds of efforts, so statisticians are 22 going to kill you, if you don't randomize by 185 1 community. 2 You can't get a school-based intervention 3 through any NIH-based funding mechanism. I don't 4 see how you could do it, other than by communities, 5 in here. 6 MR. CARLSON: Just a followup: How do 7 you deal with the problem that, in all likelihood, 8 since we know the number that's available, the number 9 of communities that we can land-mine and the period 10 is going to be very small, so you won't have the 11 problem of clustering; it will be such a small 12 sample. 13 I don't know that we can afford 20, 14 because 20 communities, randomly assigned, is an 15 extremely expensive undertaking. That's more than 16 $20 million. 17 You take one community or two, and you 18 can do it however you want to. Statisticians fight 19 about this all the time, but you have within the 20 community, thousands of participants, and you 21 randomly assign within that community, half of them 22 incented, half of them not. 186 1 The point right here, why do they talk to 2 each other? That happens all the time. The point of 3 this is to have an added benefit. You're not taking 4 anything away from them. 5 You can't afford to give it to everybody. 6 DR. GLANZ: I think we have more language 7 difficulties here. "Community" can be very different 8 things. You can be talking about a small town. 9 Statistically speaking, it's more the 10 number of communities than the number of people 11 within a community. It's not randomized, like New 12 York and Chicago. That's not going to be doable. 13 As far as the kind of background noise, 14 whether it's coming from the people signed up for 15 WIC or at school, there are ways to deal with that, 16 where you can identify those things, where you're 17 electing for the communities to come in. 18 That's what I think. 19 DR. TOWNSEND: A couple of thoughts -- 20 if the decision is in concrete, to go with the 21 randomized trial, and we find that we have to have 22 at least 20 communities and you can't afford it, and 187 1 there's the possibility of randomly selecting 2 households within each community, if that becomes a 3 no-go, because of the other issues of equitability, 4 that brings us back to a prospective longitudinal 5 design that's much cheaper. 6 Within that, you could have many of these 7 other subsets going on. 8 MR. CARLSON: Gus? 9 MR. SCHUMACHER: Looking at the 10 evaluation language, do we also want to look at 11 evaluating the optional delivery mechanisms? 12 In California, someone mentioned this 13 morning, that they have legislation on the books, 14 where, if you could take your EBT card to a market 15 and buy fruits and vegetables during that month, you 16 get a bonus on the card added in for the next month. 17 Then there's the option of gift cards, 18 then there's the option the GAO was mentioning this 19 morning. I was talking to Admiral Cohen at lunch 20 today about some of the innovative work that's going 21 on in adding on a regular card, so much technology. 22 You can do EBT, WIC, and you can get all 188 1 kinds of things with the cards. 2 There's also paper vouchers, but will you 3 will be looking at those options and then evaluating 4 them, is the question. 5 MR. CARLSON: That is a question. In an 6 ideal world, if I were to have unlimited resources 7 to design this, I'd be thinking about some sort of 8 randomized design, incorporating different levels of 9 incentives, the forms of the incentives, different 10 communities, yadah, yadah, yadah. 11 No one has yet found the ideal world, so 12 we have to make some choices. The choices are going 13 to be driven by what the most likely combination for 14 success is going to be, and we're almost certain to 15 be wrong on that. 16 MS. FOERSTER: To pick up on Marilyn's 17 comment, there is technology now, but in terms of a 18 longitudinal design, we had a presentation a few 19 weeks ago from an organization -- I can't remember 20 the name of it, but where they will give households 21 who don't have them, computers. Those who have the 22 computers, sign up, and you buy questions. 189 1 So, you can track, you can do samples of 2 low-income, Latino, and so forth, households, so that 3 you can do this kind of work, very economically, in a 4 very short period of time. 5 The old fashioned approach of having to 6 establish a center and that sort of thing, can be 7 overcome with technology, using some of the tools 8 that have been discussed, of computer delivery. 9 It may be simplified, but I think there 10 are some other options. 11 DR. HERSEY: I guess the reason that I'm 12 on this panel, is that my study was cited in Kathy 13 Roork's GAO report. I'm the "et al." 14 (Laughter.) 15 DR. HERSEY: When we did review those 16 studies, we were finding, in general, the median 17 effect size is about half a serving. Those aren't 18 huge samples. 19 Even in the Los Angeles study, where you 20 work with a $10 coupon, we're finding about 8/10ths, 21 almost maybe one serving extra of fruits and 22 vegetables. So that both encourages you to get the 190 1 benchmark, but it is, I guess, a small sample. 2 Now, having said that, let's review what 3 you can randomize fairly cheaply and what you can't. 4 Fairly cheaply, once you've got a randomized trial 5 within the same community, you can alter incentive 6 levels. 7 There is not much added cost, except for 8 the noise of explaining to people, life ain't fair. 9 I've also heard this morning, lots of 10 stuff about contextual, and that urban is very 11 different than rural; California is different than 12 Pennsylvania, and we could go on forever. 13 That argues much for multiple 14 communities. You've got real clustering effects, 15 and there are some of the delivery-mode questions: 16 Should I be doing this as an added benefit? You add 17 a voucher or has a coupon probably got to be 18 implemented differently in different communities and 19 then you can randomize incentive amounts within 20 that? 21 I guess I'd be playing with those kinds 22 of design tradeoffs and really thinking about, if 191 1 I'm trying to deal with those clustering effects, 2 I'm going to want 16 to 20 communities. 3 DR. BARANOWSKI: But the statistical 4 power -- the statisticians have done studies where 5 they have shown that the power, if you get above 50 6 per cluster, there's no additional benefit, so 7 either that argues that you go for small 8 communities, or that you randomly select people 9 within the community to receive the improvement. 10 MR. CARLSON: I wonder if those power 11 calculations would assume an effect size a little 12 bigger. 13 DR. BARANOWSKI: You increase the number 14 of clusters. 15 MR. CARLSON: Absolutely. 16 DR. KIRLIN: You also have to consider 17 the implications of the cost at the store level, the 18 sample size, of consumers, of participants. 19 You might just have a marginal cost 20 increase, and you'd have to hire more people, but 21 your costs for other parts of the implementation and 22 possibly of the data collection, are really fuzzy. 192 1 DR. GLANZ: One thing we haven't talked 2 about, is, do you get everybody in communities? We 3 haven't talked about households. 4 We're in a research climate where it 5 isn't what it used to be. Whether people who do not 6 get calls or actually not allowed to call, you are 7 asking people to allow access to personal data, 8 given that I think we've kind of got this consensus 9 that you can't do this, just by studying data from 10 stores and checkouts. 11 That has to be factored into anything. 12 You might randomly select a household within a 13 community, but in the community, you're going to 14 have problems with some households, especially if 15 you're trying to do this in a timely way. 16 We get access to people in healthcare 17 plans, through their healthcare plans. We have 18 access to claims data and they still are responding 19 a year later to the question, will you participate 20 in the study? 21 So there's a lot going on here. Twenty, 22 30, or 60 percent. 193 1 DR. DEVANEY: Remember, this is in the 2 context of intervention and benefits that they 3 receive from day one. That's a lot different, in 4 terms -- 5 DR. GLANZ: They're going to have to be 6 willing to be in the study. 7 DR. DEVANEY: Absolutely. I do think 8 that there's a little bit -- there's different 9 objectives. I want to remind you that what we want 10 to do, is learn from the pilot. 11 Standardizing across 20 communities, is 12 not a community intervention, it's an individual 13 intervention, and in those individual households, 14 the question is being asked, whether it's going to 15 change their consumption, not whether it's going to 16 change their community. 17 The next step is to get ten more in the 18 pilot and then go larger, and then I absolutely 19 agree that we have to undergo randomization, but 20 this is a first step. 21 MR. CARLSON: Other thoughts on that? I 22 have one other question. 194 1 MS. FOERSTER: The only response I would 2 have, is that we believe that people are a function 3 of their peer group, their community, the larger 4 social environment. I think that if you narrow it 5 down to saying, okay, what would a ten, 20, or 30 or 6 40-percent incentive do, and ignore the other 7 aspects, then you probably are going to get a much 8 lower result than you would, if you really did think 9 of it in a broader perspective. 10 And if our objective is to try to change 11 diets, or to try to change the composition of the 12 grocery basket, I think you really do need to think 13 about it much more broadly than what you're saying. 14 I understand what you're saying. 15 DR. DEVANEY: Your point is exactly what 16 I suggested; that they want to stay within the 17 community. They are the context, their behavior is 18 the product of everything else that's going on. 19 So, if we go across 20 communities, how 20 do we isolate the effects of the difference or the 21 variations among those communities? 22 So, within the community, it's fixed. 195 1 When you take those who live within that particular 2 area, some of them are going to be benefit, some of 3 them not. You will have the ability to learn more. 4 DR. GLANZ: I just want to add one thing; 5 there is a cost tradeoff at the community level, 6 aside from this, the one that Barbara brought up. 7 You're going to be collecting store data, 8 sales data, et cetera. That is going to be done at 9 the community level. 10 And with X-number of communities, there's 11 going to be an economy of scale and it can be studied 12 that way. 13 Clearly, if we take as a given, that 14 randomization would occur within a community, how 15 many communities would you want to include in the 16 study? 17 DR. DEVANEY: I depends on how far the 18 $20 million will go. I mean, it's ideal not to put 19 all your eggs in one basket, in New York City or 20 whatever. 21 You want to have some geographic 22 representation, but I think, within the separate 196 1 pilots, I don't think they're going to aggregate 2 across all communities. It's going to be within 3 them, that you would do the pilot. In each 4 community, you get what you want at the community 5 level, then you can have a range of impacts across 6 the different communities. 7 I don't know how much $20 million buys. 8 MR. CARLSON: That was actually a 9 precedent. In the rollout of EBT, we started in one 10 small pilot in Reading, Pennsylvania, just to see if 11 the technology would work, and we rolled it out in 12 five other states, to see if something was unique 13 about Reading. 14 And then we rolled it out for the entire 15 state, Maryland. Then we started pushing it out 16 there, and that was an incremental approach there. 17 Marilyn, you may have the last word 18 there, before I open it up. Make it a good one. 19 DR. TOWNSEND: In listening to all these 20 comments and so many research questions, I guess I 21 want to really reinforce what Steven just said about 22 the incremental approach. 197 1 There are so many research questions that 2 could and need to be answered, before we ever get to 3 these trials, the randomized trials. Once we get 4 there, we're talking about 20 communities. 5 That N of 20, only addresses one measure. 6 So we have multiple questions. 7 MR. CARLSON: Have you all been listening 8 to the series on NPR about multitasking, how we 9 think we can do it? It's not so good. 10 DR. TOWNSEND: I guess we need to 11 multitask. 12 MR. CARLSON: Thank you all for your 13 comments so far. Be prepared. Same rules as before. 14 The mike is open. Please identify your name and 15 affiliation. 16 DR. PALMER: Deb Palmer, from Rutgers. 17 You've already heard from me. I have three 18 comments/questions. 19 The first one, real quick, is, one 20 location, in terms of regionalization. The other 21 thing I heard a lot of people saying things about, 22 was, why did I get to do it and this one didn't, et 198 1 cetera, et cetera. 2 I think we need to be a little bit more 3 blunt. They're really used to going to McDonald's 4 and pulling the thing off, and they maybe or maybe 5 did not win a cup of coffee. They're used to 6 getting their cold cards in the mail. They scratch 7 this off and you get 30 percent off. 8 I don't know why we couldn't take everyone 9 in the same who's getting SNAP benefits, to scratch 10 this off and you might get 30 percent off. 11 But my question that I want to put to the 12 panel, is this: It's my understanding that the 13 measurement of certain markers for increased fruit 14 and vegetable intake, are far less intrusive than 15 they used to be, things like collecting saliva, et 16 cetera. 17 I think that smaller samples, with 18 greater impact or power, would make more sense than 19 larger samples. I know that we're a survey, a 20 national survey group in here, but I just wondered 21 about that no one mentioned doing any kind of 22 physiological testing. 199 1 I think that's the bottom line; we want 2 to know, are they eating more fruits and vegetables 3 or not, and I don't think -- I think, probably, the 4 best way to go, which is what we're doing this for -- 5 and they might lose money later, if they don't get 6 the right answer -- but shopping data, I think, in 7 part, would help. 8 But, again, it was brought up that many 9 people purchase fruits and vegetables outside the 10 grocery store, so I don't know that that makes 11 sense. 12 As Marilyn said, if we did look at a 13 physiological measure, we've really cut down on that 14 level, but if any of you can answer that -- none of 15 you mentioned physiological measures. 16 DR. TOWNSEND: Could I just say something 17 quickly? To validate a fruit and vegetable tool, we 18 did select peer measures from data in the community 19 about ten years ago. 20 Nowadays, that probably would be very 21 difficult. I haven't done the saliva, so I don't 22 know about that, but the best part about a 200 1 biomarker, is that it's not self-reporting. 2 DR. BARANOWSKI: There are two kinds of 3 biomarkers: One kind -- one is related, and there 4 are others that are kind of related, but subject to 5 variation by a lot of factors, and they're terribly 6 imperfect. 7 The major reason we stay away from 8 biomarkers, is that they don't like that. 9 (Laughter.) 10 DR. BARANOWSKI: We did a small group of 11 tolerance tests, which are substantial tests. We 12 had a huge campaign that we waged, 4400. 13 My guess is, we're going to have to have 14 an incentive. I don't know of any saliva test 15 that's related to fruit and vegetable intake. 16 Susan Main, is her name, at Yale, who's 17 got a grant to look at applying infrared technology 18 to skin. Apparently, the amount of Vitamin A in 19 your system, shows up in the skin. 20 She's applying infrared technology to 21 this kind of thing, and that's the only technology 22 I'm aware of, that shows this. I haven't seen any 201 1 validation in the literature that I read. 2 MR. CARLSON: Yes, sir. 3 MR. WENNING: Tom Wenning, with the 4 National Grocers Association. Kathy just left the 5 room and I won't be able to reference her, but at 6 our meeting at GAO, there was a great deal of 7 discussion, when we were together, about how you 8 would measure the baseline. 9 It's a challenge that Congress gave, in 10 developing, let's say fruits and vegetables. How do 11 you measure whether or not -- they say the criteria 12 is, you have to show not only the nutritional 13 benefits, but the effect on obesity on both adults 14 and children. How are you going to establish the 15 baseline to show this? 16 DR. GLANZ: That's a good question. As I 17 understand it, it's not intended to measure the 18 impact. I didn't get any kind of timeframe that 19 we're talking about. 20 It's a greatly increased intake of fruit 21 and vegetable consumption, replacing more high- 22 calorie, high-fat foods. I don't think we can do 202 1 that, even if the expectation is that we can do that. 2 The expectation would be that we'd have a 3 background of literature that talks about fruits and 4 vegetables, to see what the impact is on weight. 5 DR. DEVANEY: The beauty of a randomized 6 control, whether in the community or the individual, 7 is that the control group is essentially the 8 baseline. You don't need a baseline when you have a 9 randomized control. 10 It often helps to have a baseline, because 11 you increase your power by having the baseline, but 12 it's not necessary. 13 MS. SANTORO: Ellen Santoro, Catalina 14 Marketing. 15 I do believe we can replicate it with a 16 double-blind placebo study, within our system. What 17 we can do, is select information on the shopper. 18 We identify those in the program, and what 19 we can do, is create to match households a matched 20 household panel, store by store. We're showing the 21 community at the store level, we're controlling for 22 price and the changing price situation, the changes 203 1 in the availability of various types of products 2 within the store, and other economic factors. 3 All of that variation, we can control for 4 and we can do this in a very economic manner, store- 5 by-store, to create these matched households, and we 6 can do that across numerous stores throughout the 7 country, and, therefore, get that level of geographic 8 representation that we're looking for. 9 Importantly, we can measure the immediate 10 response, as well as the longer-term response. 11 There's no issues in terms of fraud, and we know who 12 that went to, when it went to them, if it went to 13 them, et cetera, et cetera. 14 And, importantly, we can learn, is price 15 really the barrier here, or a combination of price 16 and something else? 17 Ultimately, what is the change in that 18 market basket? Then, importantly, longer-term, 19 perhaps after the program is over, what do we see in 20 terms of ultimately changing behavior? 21 We might find that the price incentive, 22 alone, is not something, and, therefore, we just 204 1 kind of modified, learned, and moved on. 2 I think it's just the most economic and 3 well-controlled approach. 4 MS. FOERSTER: Can I ask how that would 5 interface with EBT? It's a different proprietary 6 system. 7 Would there be a computerized linkage, so 8 that you would know that, for example, at the end of 9 the month, when people aren't eating from EBT, that 10 they are still SNAP customers? How would that work? 11 MS. SANTORO: Certainly, we have respect 12 for all the privacy issues, but if we have an EBT 13 identifier, in addition to a pretty good shopper 14 card, we could link to those and then track that 15 individual household over time. 16 MR. SCHUMACHER: If they go to the same 17 store. 18 MS. SANTORO: What I was saying, is, we 19 look store-by-store and select those stores that are 20 geographically ideal. 21 In addition, we can look at -- we 22 basically have perfect data behavior in terms of 205 1 what was purchased, and we can control for all of 2 those factors, as long as you know how often that 3 person shops in that specific store, what they 4 purchased in each and every shopping occasion. 5 The quality of the data is so robust. 6 You've already looked at the analytic approach and 7 the design of the study, so you'd be amazed at how 8 well controlled the study is. 9 DR. DEVANEY: Your data actually are 10 linked to the SNAP records that show whether this 11 household got the financial incentive? 12 MS. SANTORO: SNAP is, I believe, treated. 13 MR. CARLSON: The fact that they have to 14 swipe the key card. As I understood the discussion 15 earlier, when they're swiping the key card in the 16 Catalina system or the store system, it picks up the 17 identifier on the card. 18 DR. DEVANEY: From there, you can tell 19 whether they received the incentive or not. 20 MR. CARLSON: Presumably, the incentive 21 is triggered through the EBT card. 22 MS. SANTORO: Additionally, we do work 206 1 with consumers, on the survey. If they're not 2 buying it, they're not consciously, incrementally 3 eating fruits and vegetables, if they're not buying 4 it, we have not been successful. 5 Certainly, in terms of follow-up, if they 6 buy it -- and, importantly, are we changing their 7 attitudes? 8 The average tenure on the program, is 9 about nine months. We don't expect to see much 10 effecto on their behavior over that short term. 11 We check behavior markers, such as weight 12 reduction, and those are ultimately impacted. Then 13 we could, ultimately, if we had this very controlled 14 environment, measure, longer-term, things like weight 15 reduction. 16 Importantly, it will be very different, 17 if you have children who eat what you prepare for 18 them, versus children who are teenagers, who kind of 19 decide what they're going to eat. 20 Importantly, we can look at it in terms 21 of understanding which communities and which types 22 of household profiles will, in fact, be able to be 207 1 ultimately influenced. 2 MR. CARLSON: If I could make one 3 clarification, the median length of participation 4 among new entrants, is roughly eight or nine months. 5 There are a large number of folks who 6 stay much longer than eight or nine months, but the 7 average duration, the cross section, is more on the 8 order of years. 9 If you're applying for food stamps for 10 the first time, the chances are, you'll be gone 11 within eight months, and the chances are, you won't 12 come back. 13 Jessica? 14 MS. SHAHID: I had one question, and it 15 may actually be more for the next panel. My 16 question is, are you saying that you can grab this 17 data for evaluative purposes, because it's an EBT 18 card that's used to make the purchase? 19 But many states use various programs as 20 part of their EBT card. They might be using that 21 cash on their EBT card; they might be using their 22 food benefits on the EBT card. 208 1 Do you also distinguish that it was SNAP 2 that made that purchase, or is it only an EBT 3 recognition? Does that make sense, in terms of my 4 question? 5 MS. SANTORO: That's something we have to 6 evaluate. We have not yet done this program for the 7 SNAP ed. 8 MS. SHAHID: WIC and EBT aren't are the 9 same card in California. My question is, is it 10 simply EBT, or is it EBT broken down? 11 It's a different stream of funding. All 12 of you guys are from Catalina? 13 (Laughter.) 14 MS. SHAHID: It is not simply the EBT; it 15 is the tender-type. It is the EBT that can be then 16 be broken into a subgroup; that's all I needed to 17 know; thank you. 18 DR. TOWNSEND: Catalina can also look at 19 the debit card use when the food stamp money has 20 ended. They can bring together, the debit card to 21 the SNAP EBT card and make that one purchase. 22 What you can't do, is track cash. 209 1 DR. KIRLIN: They use the loyalty card. 2 MR. CARLSON: Right. 3 DR. TOWNSEND: If they use the loyalty 4 card, they could track cash, but, with or without 5 the loyalty card, the can track the debit card. 6 Isn't that what you said? 7 MR. BALMER: We're actually here to help, 8 if we have some solutions for you all with reference 9 to this issue. What we do see, is, we see tender- 10 type, so when we see a transaction, we basically 11 overlay a retailer's point-of-sale. 12 If a consumer insists that we use some 13 form of tender, a key card and a debit card, we can 14 link those two things together. We don't know that 15 it's a household that has a certain type of name or 16 address. 17 Consistently, we see this EBT card with 18 this other tender card, so we link all the purchase 19 behaviors. 20 MR. CARLSON: Is that in instances where 21 they use the EBT and the debit card at the same 22 sale? 210 1 MR. BALMER: Correct. 2 MR. CARLSON: But if you use the EBT card 3 today and use your debit card tomorrow, there's no 4 way of knowing, except in this case, if they've done 5 it before frequently enough? 6 MR. BALMER: We need to look at this a 7 little more thoroughly. That's why we just got the 8 numbers, but we feel there would be the opportunity 9 for us to see the EBT card and a debit card. 10 If we see that debit card again, you can 11 link all of those purchases together. 12 MR. CARLSON: That clarifies it. We have 13 two more. If there's any possibility, let's make 14 them brief. 15 MR. THOMASES: Ben Thomases. So, 16 listening the panel this morning, I heard, I think, 17 the consensus that the program, when it is designed 18 to be most impactful, would combine the incentive 19 with something that might be called nutritional 20 education or cultural marketing or something, 21 changing the kind of way of the people that make 22 these decisions. 211 1 I wasn't sure you folks were saying -- 2 when you were talking about contamination, that you 3 felt that that would make the program difficult or 4 impossible to evaluate, when combined with some 5 other type of effort. 6 DR. GLANZ: I don't think that's what we 7 were saying. I think that what we were saying, is 8 that some of those things are going on, anyway, not 9 necessarily related to the SNAP program or 10 households. 11 They might just be going around to some 12 places and not others. 13 DR. BARANOWSKI: There is lots of 14 interest in whether the incentive program, by 15 itself, would have an impact without the 16 intervention program, and so I would like to see a 17 two-factor design, where there's an incentive 18 program, there's an education program, so there's a 19 group that gets nothing, there's a group that gets 20 one, a group that gets the other, then a group that 21 gets both, so it's a four-armed trial. 22 You can then systematically evaluate the 212 1 contribution of the incentive system, versus the 2 education system, in impacting the dietary intakes 3 of the participants. 4 DR. KIRLIN: I would add, if you're going 5 to do that at the intervention, that the nutritional 6 education intervention that you have, be something 7 that you see continuing down the road. It's not 8 worth testing, if it's not implementable long-term. 9 MS. FOERSTER: By that same factor, is 10 the incentive going to be available down the road, 11 too? So, both of those questions have to be 12 answered, and are we really trying to increase the 13 value of the food benefit, or are we trying, for 14 people that go off the program, we want them to 15 continue to have healthy food practices? You have 16 to think long-term. 17 DR. BARANOWSKI: That argues for two post 18 assessments. There should be one assessment right at 19 the end of the incentive program, to look at what the 20 short-term effect is. 21 Once you remove the incentive, you then 22 want another assessment, some equal period. So, if 213 1 the intervention is six months, the second 2 assessment would be 12 months. 3 So, looking at what the maintenance of 4 the effect is over a longer period of time, that's 5 exactly why you want to start measuring the 6 intrinsic and extrinsic motivation issues, because 7 all the extrinsic motivation literature, would 8 suggest that that behavior is going to go completely 9 away, once you remove the incentive. 10 So it's a critical issue, and you need a 11 separate assessment, after the end of the 12 intervention. 13 MR. CARLSON: Yes, ma'am? 14 MS. WINCH: Alisa Winch, with the 15 Congressional Hunger Center. How do you recommend 16 involving SNAP participants in the design and 17 evaluation of these incentive programs? 18 MR. PEREIRA: I'm actually not that 19 familiar with that program, and I'm probably a bad 20 person to answer the question, but it's a very good 21 point. 22 It was raised earlier this morning, that 214 1 all of the pieces that get somebody to respond to 2 this program, and help them eating better the food, 3 the relevant issue, I think, speaks to talking to 4 people who are actually participants in the program, 5 who understand how do you relevantly communicate with 6 them. 7 There's a lot we can do through studies 8 and analytics, understanding behavior and looking at 9 issues of relevancy. In our opinion, a lot of the 10 work we do, is all around gearing to relevancy, with 11 the consumer, in this case, from a slightly 12 different angle. 13 MR. CARLSON: I would just add to that, 14 that we have a fairly long history of doing a lot of 15 informative research, especially on education and 16 promotion activities. Our method is to learn what 17 works with them and what motivates them, and so 18 forth. 19 One of the questions earlier this 20 morning, was, what do we need to do in order to 21 inform the recipients that there is this incentive, 22 so I think your point is well taken, and it's 215 1 consistent with a notion that's come up here several 2 times, that maybe one of the early steps, is a series 3 of small steps, to figure out what works and what 4 might not. 5 It's always, in the end, a question of 6 balance. The more small studies we do in the next 7 year or two, the longer it is, before we get to the 8 granddaddy of all pilots. 9 I think we do feel a little pressure to 10 have this done before the next one is taken up, 11 which is, god forbid, only four years from now. 12 (Laughter.) 13 MR. CARLSON: I may not be around for 14 that one, but someone will be. How much we can do 15 in the interim, will be a difficult thing to 16 determine. 17 Once again, let me thank our panel. You 18 did a marvelous job, and it was very stimulating. 19 You gave us some ideas. Thank you very much. 20 We're going to take a 15-minute break. 21 By this clock out here -- we'll make it a 13-minute 22 break. By this clock in the center, it will be ten 216 1 of 3:00. 2 (Recess.) 3 MR. CARLSON: We're on the home stretch. 4 It's been a long day, I understand. We're now ready 5 for our third and final panel session. I was asked - 6 - I recognize that the day is growing long, and you 7 all have things scheduled, and not all of you may be 8 able to stick around till the very end. We'll try to 9 wrap up. 10 I was asked to at least mention the 11 general sense of timelines. What happens today? 12 Beyond being evasive, we are thinking at the moment, 13 that it's important to the ultimate success of this 14 project, to launch ourselves down the right path, and 15 that's going to take us some time. 16 We had some discussions about moving into 17 the field very rapidly, and I think we've pretty much 18 concluded that, in the long run, the project will 19 benefit from a more thoughtful, careful consideration 20 of the options available to us. 21 With the kinds of commentary we've heard 22 today, everything is subject to change, but, at the 217 1 moment, we're probably working on a schedule that 2 will see the release of a solicitation for 3 applications for demonstration sites and for 4 independent evaluation, probably -- if I say this, 5 Carole is going to kill me -- I hope we'll get that 6 out in ten months. 7 Carole would probably like to get it out 8 in 12 months, but it is probably a ways away, before 9 we actually launch. 10 Again, if we take to heart, some of the 11 advice we got today, that we ought to do a larger 12 number of formative studies, maybe that happens more 13 quickly. That had not been factored into our 14 original thinking. 15 All I can advise you, those of you who 16 are interested, either in participating in the 17 demonstrations or the evaluations, or just staying 18 informed about what's going on, stay tuned and we'll 19 let you know. 20 After talking about the nutritional 21 framework of providing incentives, after talking 22 about how to put ourselves in a position to 218 1 understand whether it worked, the panel before us 2 this afternoon, has been asked to deal with the 3 question of how do we make it work. How do we 4 actually put this on the ground? 5 We, again, have a wide range of folks to 6 share their views, opinions, and thoughts with us. 7 I'm going to go back to the routine and start with 8 this end of the table, in reverse alphabetical 9 order, almost. It's a little whacked out at that 10 end. 11 Here we have Jan Walters, an EBT Project 12 Director from the Iowa Department of Human Services - 13 - not surprisingly, Iowa. 14 Next to her, Tom Wenning, the Sr. Vice 15 President and General Counsel of the National 16 Grocers Association. 17 Next to him, we have Mitch Klein, whom 18 you've heard from, who is the Vice President of 19 Government Relations and Retail Services at Krasdale 20 Foods, operating out of White Plains, New York. 21 Next to him, Peter Relich, Sr. Vice 22 President at Maximus. We had a short conversation 219 1 before this, where he reminded me that he did some 2 recent work on EBT and food stamps and alternative 3 assessments, and had some connections to what we've 4 been doing. 5 Next to him is Errol Baker, an Associate 6 with Booz, Allen, Hamilton, who has been heavily 7 engaged in helping FNS in a wide range of 8 activities. 9 Eugene Costa is the Executive Vice 10 President of Imadgen. You'll have to tell us what 11 Imadgen does. 12 MR. COSTA: Technical consulting. 13 MR. CARLSON: Shay Alon is Product 14 Manager of the Government Solutions Division of 15 EFUNDS. 16 Lastly, and certainly not least, is Dick 17 Angerosa, who is the Vice President of J.P. Morgan 18 Chase, Electronic Financial Services. 19 Again, we want to welcome all of you. 20 You all are going to quickly discover that I have 21 way less background in system features, than I 22 pretended I had in the first two discussions. You 220 1 may receive more help from my colleagues and the 2 audience, who have been working in this area for 3 awhile, but on the same plan as before, we have a 4 set of questions that we'll pose to the panel, and 5 give anyone who wants the opportunity to talk, the 6 opportunity to do so. 7 Don't feel that you have to, if you don't 8 feel you have anything to say on a particular issue, 9 but we'll sure everybody has a chance to say what 10 they need to say. There will be time at the end for 11 all those who are successful in the marathon, to get 12 to the 26th mile. 13 We'll have that opportunity for audience 14 participation, then we'll do a short wrap-up. With 15 that, the first question. I'm starting to feel like 16 a game show host. 17 (Laughter.) 18 MR. CARLSON: Panel, what are the 19 implications of choosing between an incentive that 20 assumes the form of a discount, available before 21 purchase, a discount upon purchase of fruits and 22 vegetables, or a bonus afterwards? 221 1 This is one of the things we've had a lot 2 of conversation about. Are there some system 3 implications to structuring this to be tied to a 4 particular behavior on the part of recipients, as 5 opposed to just providing an amount of money that 6 has to be used for a specific purpose, as opposed to 7 providing them that amount of money? 8 MR. COSTA: I'll go first. Obviously, 9 this will affect the amount and location of the 10 development work that needs to be done. Certainly, 11 at a very simplistic level, you can capture a 12 transaction and give it a discount and then that's 13 all you need to do. 14 But if you want to compare against 15 balances or thresholds, before a benefit is given, 16 then that's going to require development, an 17 accumulation and aggregation of data. 18 So, starting off with that -- 19 MR. BAKER: Just quickly, I'll just add 20 to what Gene said about data being a big driver. I 21 was listening to a lot of this morning's 22 conversation, and at some point, there's going to 222 1 have to be a decision. 2 Sort of looking at output, might help 3 determine, ultimately, how we want this to look, 4 meaning, taking a look at some of what some of the 5 reporting requirements are going to be; taking a 6 look at some of those things that we're going to 7 want to measure, and use that to help drive us. 8 If it's going to be before, during, or 9 after, certainly from a systems standpoint, it's 10 just the complexity, the time and the cost, that's 11 going to be associated with each, and, certainly, 12 the discount. 13 So I think the driver in the data and a 14 lot of the system changes that might result, 15 whichever solution we end up going to. Again, I 16 think, looking at it from a reporting standpoint, 17 how complex we want the reports to be, in order to 18 track, is it going to be by, for instance, apples, 19 if you want to get to that granularity, versus a 20 fruit or vegetable. 21 If that is in the purchase, leave it at 22 that, and let's say you apply the discounts, say $2 223 1 to that, saying there's grapes and apples in there, 2 there's a different amount that goes to grapes, 3 versus apples, so you know, there has to be a 4 decision on how complex the data should be. 5 MR. CARLSON: We'll get to Dick. 6 MR. ANGEROSA: If you talk about what are 7 the system implications of setting up the solution 8 that you choose, this largely follows -- it's no 9 surprise that the easiest solution is the simplest 10 and the cheapest, but they have a lot of downsides. 11 If you get a little more complex of a 12 solution, the net benefits are much greater, but if 13 you take those three approaches, if you look at a 14 pre-purchase decision, discount, if we're talking 15 about a coupon or something, the cardholder goes 16 into the store to shop, presents that to the 17 retailer, takes the $10 healthy food purchase, and 18 immediately gets the 20-percent discount, the 19 cardholder pays $8 for his purchase, the downside to 20 that is, on the EBT host, the host data records an $8 21 food stamp purchase. 22 There's no indication that it's a healthy 224 1 food stamp purchase; there's no electronic audit 2 trail. Plus, the retailer has to go through a 3 process to redeem the coupon or vouchers, in order to 4 be paid, so it's simple and easy to implement, but in 5 terms of your reporting statistics and everything 6 else, it's really lacking. 7 If you look at a solution where there's 8 reimbursement to the client, after the fact, right 9 out the gate, we've heard a couple of times already 10 today, that they want that immediate gratification; 11 they don't want to wait. 12 If you're going to reimburse the client 13 after the fact, that presupposes that you have to be 14 able to record the fact that the client did transact 15 a healthy food purchase. So, on the inquiring side, 16 that client would be uniquely identified to the host 17 processor, so that we could identify those purchases 18 and generate, automatically, some type of credit back 19 to the client. 20 If you are going to the trouble of 21 identifying the healthy purchase to begin with, I'd 22 give the client the benefit in the checkout line, on 225 1 the spot. 2 We now come to the middle solution, which 3 is a discount at the point of sale, at the checkout. 4 In the GAO session last year, this is what we talked 5 about fairly extensively, and we discussed it today 6 during the break. 7 What would happen, is, the client will 8 walk into the store, make that same $10 healthy food 9 purchase. The acquirer would send in a transaction 10 that identifies to that host processor, that this 11 was, in fact, a purchase of healthy food. 12 Let's say there is a 20-percent discount 13 that applies. In the model we talked about, that 14 discount could be controlled by FNS. 15 So, the host would get the transaction, 16 and immediately generates a $2 credit to the 17 cardholder. Temporarily, their balance is down. 18 Say they had $10 in their account. They've got $12, 19 then process the full $10 debit, leaving the 20 cardholder with $2. 21 That credit to the account, will be in 22 the form of a deposit. 226 1 Now, the cardholder gets the benefit of 2 the discount on the spot, in the checkout line. 3 It's reflected on the receipt they get back. 4 The retailer doesn't have to do anything 5 differently, doesn't have to calculate a discount; 6 it just tells us, this is a healthy food purchase, 7 and he gets settlement for the full $10. 8 So, the client walks away, they've got 9 their discount on the spot, and this does a lot of 10 things for you. 11 The host processors have captured the 12 discrete contents of this transaction or 13 transactions, and the possibilities for statistical 14 analysis, reporting, and things like that, are 15 virtually limitless, because you now have a detailed 16 record on the host system, the host processors, of 17 which cardholder did the transaction, what store was 18 he at, how much he spent, what was the amount of the 19 discount. 20 Giving the discount in the form of 21 credit, which is a deposit, gives you a way to track 22 the screened income from day to day, as to how many 227 1 discounts they've given. 2 If you make that credit, a deposit, it 3 folds very nicely into the existing settlement 4 system. It will be included automatically in the 5 letter of credit, it can be settled just like the 6 normal food stamps, so it's not anything really 7 different that needs to be done there. 8 So, in terms of cost and complexity, 9 sure, there's a lot for the host processor to do; 10 there's a lot of software to develop, but what you 11 end up with, is a much more robust system, with 12 benefits for the cardholder, who gets the discount 13 on the spot; benefits to the retailer, who gets one 14 payment through their own settlement process; and 15 all the statistical data you might eventually need. 16 MR. CARLSON: Point of clarification. 17 With the process you just described, would it 18 require the clerk at the checkout counter, to 19 separate out, as part of the transaction, the charge 20 of the foods that are being discounted or rewarded 21 here? 22 MR. ANGEROSA: Obviously, the more 228 1 sophisticated retailer with the UPC codes, in the 2 checkout, they could analyze the UPC for healthy 3 foods. 4 For the smaller retailer, they'd have to 5 separate and figure it out. 6 I didn't want to over-complicate the 7 computer. I was talking about this before. There 8 has been a change already made, based on the fact 9 that we were asked to look into this a year or two 10 ago by the State of California. We actually had 11 extensive meetings with them, in finding a solution. 12 There's already a change to the X-9 13 standard to identify healthy food purchase. That 14 one piece of data can be used in conjunction with 15 the existing data elements, so that that a 16 nutritional food purchase can come down by itself or 17 as part of an overall food stamp purchase. 18 If you think about it, at the checkout, 19 it's maybe one more step, a final swipe of the card, 20 under the PIN, how you're going to pay for these food 21 stamps, what's the amount of purchase. 22 One more step; what's the amount of the 229 1 nutritional item that's part of this purchase? That 2 can all be done in the same message. 3 MS. WALTERS: In Iowa, we have worked 4 really hard the last six years, on restoring access 5 for our food assistance recipients, to be able to go 6 the farmers markets to shop. I agree with Dick, on 7 the idea that it would work great for a brick-and- 8 mortar type retailer, however, that will not work for 9 farmers markets. 10 When you look back at why the food stamp 11 program was formed, one of the main underwriting 12 principles, was to help increase the income and 13 revenue stream for farmers. 14 So when you design any pilot or any 15 disincentive, I would beg people to please consider 16 how it will work for a farmers market. You look at 17 the point-of-sale equipment that farmers are using 18 today, whether it be a token process or whether 19 farmers have individual machines, today, there is no 20 way for that equipment to differentiate whether 21 something is a healthy food item or a non-healthy. 22 Food assistance recipients today, can go 230 1 buy bread and meat, eggs, cheese, et cetera, but 2 that doesn't fall into your fruits and vegetables 3 category, so there has to be some development or 4 some change for that equipment to include those 5 types of retailers in this pilot design. 6 MR. WENNING: I'd just like to follow up 7 on what Jan had to say. I represent the National 8 Grocers Association, which represents primarily 9 privately-held and operated retailers, who operate 10 anywhere from a single store or supermarket, to 11 maybe a couple of hundred-store chains. 12 Part of looking at it from a retailer's 13 perspective and backing up, I think, the real 14 question is, what is going to be the degree of 15 specificity that would be required in going into a 16 permanent program? 17 Is this going to be something that we're 18 going to have to identify all the different 19 products, whether they're canned, frozen, fresh; and 20 how are they going to go through the check stand? 21 And, when they go through the check stand, are they 22 going to be picked up on the UPC label? Where is 231 1 that database going to reside? 2 It's going to have to reside, either 3 centrally, through something that FNS is going to 4 identify, that these are acceptable products, or 5 it's going to have to be something that's just a 6 generic. 7 If it's anything that's just a broad 8 category of, let's say, fresh fruits and vegetables, 9 is there going to have to be something done at the 10 front end, at the check stand, that the retailer 11 would have to do to identify that as being qualified? 12 I like your idea about how the discount 13 would be handled, but I think we also have to be 14 aware of the fact that many retailers would have to 15 go through a cost of changing their system. That is 16 something that is not factored into the pilot 17 project, at least as far as our members are 18 concerned. 19 So we would have to be looking at that, 20 and Jan and I, we talked at lunch today about the 21 things they're trying to do in Iowa, which are going 22 to take time. It's going to be a process. 232 1 So, to go from a pilot program to a 2 nationwide program, much like we did with EBT, is 3 going to be a very complex process for us to go 4 through with retailers. 5 l think Mitch has his own view. 6 MR. KLEIN: Sure. We operate in seven 7 states and service approximately 3900 retailers. 8 New York is obviously our largest state. 9 The one thing about our retailers, is, if 10 we have 3900, there's 4200 that are unique. 11 What I would caution the group about, is 12 that what works in maybe one particular locale, one 13 particular area, is not going to be consistently 14 approved or working in another particular area. 15 Whether a vendor works well in one 16 particular area, may not necessarily work in another 17 part. 18 New York is a rather large state and it 19 would be a rather large, complex thing to change. 20 Most of the stores in New York City, operate on 21 state-operated equipment. 22 This is equipment that the state has 233 1 provided at no cost to them. While the J.P Morgan 2 program, which is the one who are vending that 3 equipment, is a very good system, there are 4 considerable changes you would need to make to get 5 that type of information. 6 The current state equipment gives you a 7 transaction dollar back; it doesn't tell you whether 8 I bought beer or eggs or cheese or meat or milk, you 9 get a dollar transaction back. 10 Also, if you're going to allow, 11 especially in New York City, where you are operating 12 an independent market, there are no such things as 13 chains where they're pulling back data and there's 14 accumulated data. 15 You're going to have a clerk making a 16 decision, sometimes at the register, on whether or 17 not this item may or may not be a healthy item. 18 So, it goes further into the conversation 19 down the road, on different things like that. You 20 take a lot of the control, or, as they were talking 21 about before, you can start to build a lot more bells 22 and whistles into the system, but it starts to become 234 1 very, very costly. 2 You have to have, somewhere, a budget, 3 before you just start to roll down this road, to see 4 exactly what you want to spend. 5 MR. RELICH: These are all interesting 6 comments. The part about buying beer, I like that, 7 with the food stamp program. 8 (Laughter.) 9 MR. RELICH: I guess the first thing is, 10 when we look at it, we need, I think, to separate 11 the pilot from the production, because they are two 12 different things. 13 The pilot is trying to get different 14 types of information, and, actually, it was very 15 interesting, after hearing Catalina talk. I'll call 16 it a commercial, but it was very interesting. 17 There's a lot of potential there. The 18 problem that you have with that, is that you're 19 going to be limiting yourself to only certain types 20 of stores, the ones that are using Catalina's 21 services. 22 Some other smaller grocers, aren't, while 235 1 the big ones are. It's very interesting. I looked 2 at the website while they talked, and it was very 3 interesting. 4 I think what Wenn talked about, about 5 when JP talked about what you could do with the 6 terminal, there is potential there. Jan, we talked 7 about it, and part of it is, when we look at the 8 food stamp program now, it's pretty much self- 9 regulated. 10 The retailer actually decide what is a 11 food item, what's a non-food item. A lot of them 12 have systems that do that, but in the bodegas and 13 the other ones, the retailers decide; beer, that's a 14 food item, but that's not covered by food stamps; 15 it's not a food item, and the other stuff. 16 So they break it out. The same thing 17 could happen with the SNAP program's healthy 18 incentives, but it has to be very clear. That's 19 going to be an issue. 20 I listened to the first panel talk about 21 it. I never really got a clear definition about 22 what's a healthy item. Processed catsup is a 236 1 healthy item? 2 So, once you decide that, it's crystal 3 clear, if it's all fresh fruits and vegetables, if 4 it's frozen, as long as there's no sugar or 5 whatever, that's good. That will help. 6 And then you're going to have the thing 7 with the farmers markets. Most of it is fresh 8 fruits and vegetables and a pie or something else or 9 a side of beef. 10 I don't know if a side of beef is 11 considered healthy. 12 (Laughter.) 13 MR. CARLSON: I probably should have 14 mentioned, as part of the ground rules, that catsup 15 is a vegetable. 16 (Laughter.) 17 MR. CARLSON: Let me push this a little 18 bit. The panel is anticipating a couple of 19 questions here. This is sort of a two-part 20 question. 21 One is, how important is it to have a 22 pilot operate throughout the entire range of 237 1 authorized retail stores? If it is important, under 2 what circumstances is it possible or imaginable to 3 actually include that full range of stores, in terms 4 of the technologies that are available to make it all 5 happen? 6 MR. RELICH: I've got the mike, so I'll 7 speak first. I guess that's a question that a 8 statistician should have answered. 9 From my perspective, you know, if you're 10 looking at consumer behavior, it doesn't really make 11 that much difference. Across the full spectrum of 12 retailers, what you're trying to do, is to get a good 13 cross section and measure risky behavior. 14 If you can get a good cross section of 15 those retailers where you can get that information, 16 like the retailers used by Catalina, that would be 17 great. 18 MR. ALON: The Catalina model is that if 19 you use a market, you may not purchase something in 20 the same place. 21 You may purchase your fruits and 22 vegetables in another small stand. If you're 238 1 looking just at the numbers, you may not get that. 2 That's why it has to be brought in that the best 3 solution would be to have a small state. 4 MS. WALTERS: Like Iowa? 5 (Laughter.) 6 MR. KLEIN: I think you want to have 7 reasonable cross section. I definitely think you 8 would want to look at something like the inner city. 9 The trick in the inner city, though, is, 10 people like IRA, Nielsen, or even Catalina, are not 11 really represented in the inner city, when it comes 12 to tracking those types of retailers that we have. 13 I've got 3900 stores. I have very few of 14 those companies representing any of those retailers. 15 I would say that I'm probably the largest 16 food vendor in the State of New York, and pretty 17 close to being the largest food stamp vendor. They 18 become a challenge. 19 There's nothing wrong with those 20 companies. These are small, independent operators 21 that are just outside of the radar when it comes to 22 a chain or a large operating store. 239 1 I think that if you really want to look 2 at healthy eating habits and you want to talk about 3 the inner city people, where there are issues of 4 obesity, diabetes, and not eating fruits and 5 vegetables, you want to be able to find a way to 6 focus on what their eating habits are, and on how to 7 get accurate, reliable information. 8 MR. BAKER: I think it is important that 9 you have a nice cross section of the community. You 10 want to make sure you're having as much impact as you 11 can. 12 Having said that, it's important, in 13 whatever pilot, that there needs be a look across 14 all the different types of retailers, and understand 15 some of the challenges that they might face. 16 So it think it's important that you look 17 across all different retailers, not that you're 18 going to come up with a solution that's going to 19 satisfy every need, but you'll at least get an idea 20 of the compromises that will have to be made for 21 those retailers and all participants. 22 MR. WENNING: I agree with what he said. 240 1 The other thing to what Jan was saying, whether you 2 take Iowa, I think you also have to have a rural 3 market. 4 If you took Iowa, and you said DeMoines, 5 and then there are some small towns where you'd want 6 to see what the impact is in rural communities, I 7 agree that you have to go across the whole spectrum. 8 I'm looking ahead to some of the 9 questions that you have coming up about integrity in 10 the program. You're going to run into different 11 types of abuses, depending on the type of retailer 12 you have, and you want to evaluate it and study it. 13 It would call for all types of retailers. 14 MR. CARLSON: As we're passing the mike, 15 can you imagine ways in which a system like the one 16 that Dick described -- Jan sort of said that it's 17 not workable in the farmers markets, but can you 18 imagine solutions across the wide variety of 19 retailers that are out there? 20 Rolling EBT out, was challenging, because 21 many of the stores weren't prepared. 22 MR. ANGEROSA: I was just going to talk 241 1 about that. In terms of implementing it across a 2 wide spectrum of retailers, you have to keep in mind 3 that the ability for the retailers to capture this 4 transaction on the host processor, that is, a 5 healthy purchase, is a function of software and a 6 PLS terminal. 7 If the software in the PLS terminal 8 supports that, any retailer that has a PLS terminal 9 -- and that includes a farmers market -- can process 10 these types of transactions. 11 I'll just mention a couple of times to 12 address that concern somewhat. I'm certainly not 13 trying to trivialize it. 14 Jan is correct; there are software 15 modifications necessary for the terminal load, in 16 order to make it work. There is work to be done on 17 the retailer's side. 18 A couple of things mitigate that: 19 Without going into great technical detail, we did 20 look into the solution. We tried to come up with 21 the methodology that would enable us to identify 22 those transactions with a very minimal change, as 242 1 minimal as we could possibly think necessary. 2 The other thing is that each retailer 3 does not have its own terminal or its own software. 4 There are third-party processors and different 5 groups that are terminal drivers, and one terminal 6 driver may load the software to thousands of 7 terminals, but even in the smaller stores that were 8 referred to, in New York or in the neighborhoods, 9 those people have point-of-sale terminals. 10 They are part of some EOS network. They 11 work with some of the platforms, and, yes, the 12 acquiring platform would have to change software, 13 but it's done on a large scale. 14 MR. CARLSON: While we're switching, that 15 raises another question, if you don't mind. If we're 16 talking about creating a pilot and we're operating 17 the system in a limited geographic area, is there a 18 way to make that change in those platforms for some 19 stores and not all stores? 20 MR. ANGEROSA: It's dependent on the 21 terminal. One way to control the pilot, is to only 22 load the new software with the ability to identify 243 1 the purchases in the stores you want to participate 2 in the pilot, or the areas you want to participate 3 in the pilot. 4 But remember that if you're supporting 5 this healthy food purchase system, the whole process 6 is to be ready, willing, and able to process this 7 healthy food transaction, for everybody else to send 8 in a normal food stamp transaction, just like they 9 always do -- no harm, no foul. 10 That's one we came up with this 11 controlling the area of the pilot. 12 MR. WENNING: Doesn't that come down to 13 what the nutritional panel was talking about this 14 morning? Who's going to identify what is the 15 healthy food and what's going to be on that list? 16 Is it still going to be the retailer 17 that's going to be deciding what's healthy, or is it 18 going to be USDA that's going to describe or 19 determine what products fit into that list? 20 MR. ANGEROSA: That really is a technical 21 question. From our perspective, you can have either 22 an integrated system, where the UPCs are scanned in a 244 1 large store, and, yes, they have to somehow program 2 that into their checkout software, this list of UPCs 3 for healthy food items. 4 A smaller store would scan -- would have 5 a list on the cash register, or know the guidelines 6 and say this is a healthy food and this isn't. 7 From the technical perspective of the 8 host processor, this many dollars, were healthy 9 foods, that's really a technical implementation 10 question. 11 MR. WENNING: The retailer is going to 12 have to do that. 13 MR. CARLSON: It's safe to assume that as 14 part of the pilot, the USDA will be very specific. 15 To some extent, the analogy to that, is in place 16 already and are in our regs, our requirements to 17 define what a healthy food is, and what ineligible 18 items are. 19 Because we don't put too many 20 restrictions on this, there's not any food item, 21 with a few limited exceptions, that's not clear if 22 it's a food or a chemical or controlling byproduct. 245 1 That raises some questions, but I think 2 we're likely to put out some specs as to what kinds 3 of food we're talking about, and, just as in current 4 tracking, ultimately it's the retailer's 5 responsibility to sort these out from one category 6 or another. 7 They will still meet some elements of 8 uncertainty. 9 MR. ALON: When you're doing a pilot for 10 evaluation, if you do an evaluation and you go down 11 and say, how many products you've sold and so on, 12 then you have 80 percent that are not going to have 13 information and cannot identify what is a healthy 14 food item. 15 You probably need that for the pilot. 16 MR. CARLSON: The 80 percent -- most of 17 the small stores probably don't have point-of-sale 18 scanner systems. 19 MR. ALON: The corner store, the mom- 20 and-pop, the farmers market. We have those that are 21 not using this equipment at all. 22 They have no way to identify what is a 246 1 healthy food item. 2 MR. CARLSON: There shouldn't be a whole 3 lot of use of vouchers in the SNAP program. What 4 they probably don't have, is a way of connecting the 5 EBT food purchase item, through the item they have 6 purchased. 7 In very limited circumstances, if the 8 power is out or something -- 9 MS. WALTERS: Retailers that do less than 10 $100 a month, do manual vouchers. 11 MR. CARLSON: I hope we don't have to 12 design a pilot for that. 13 (Laughter.) 14 MR. CARLSON: Let me move on. We've hit 15 on some of this, but let me sort of wrap up this 16 part of the conversation. 17 We're very proud at FNS that we've 18 encountered and dealt with a lot of challenges. 19 Can you think a little bit about, if 20 we're trying to use EBT as part of this incentive, 21 are there some obvious challenges that are going to 22 have to be addressed and some lessons that have been 247 1 learned from the rollout of EBT up to this point, to 2 inform us of how we want to proceed, the directions 3 we ought to look? 4 This is also your opportunity to talk 5 with us a little bit about what the cost would be of 6 doing all this. Are we talking about minor little 7 tweaks, little incentives? Or are they major 8 investments that would be harder for us to implement? 9 MR. COSTA: Let me start by saying that 10 for our purposes, performing this pilot, we don't 11 want to necessarily kill the golden goose. We can 12 gain a lot of efficiency in transactions, and people 13 become used to the efficiencies, so when something is 14 being developed, and it come down to terminal load, 15 certainly, if you want to do an analogous type of 16 transaction, you look at transactions with cash back. 17 When someone comes in to buy something, 18 we ask them if they want cash back with that 19 transaction, also. The same thing could be done 20 with a healthy food item, you know. 21 You're told, the food purchase is $100, 22 but that might take -- a lot of times, it's on the 248 1 recipient side of the transaction, to do cash back. 2 You know, you have to put in the 3 information on the healthy food items, so I say, 4 when you look at this, you have to look at those 5 types of lessons. 6 You don't want to remove the efficiency. 7 MR. RELICH: I'd like to say the same 8 thing. The other thing is, buyers, when they go 9 into the store, need to understand what's a healthy 10 food. 11 We talked about the 80/20, the 80 percent 12 of the retailers that don't have it automated. The 13 clerk has to be able to very quickly understand, 14 what's a healthy food, particularly. 15 Clients do care about money. They care 16 about what they're buying. They get very 17 frustrated, very quickly, if they can't identify 18 what's a healthy food item. 19 MR. BAKER: I agree. The challenge is 20 certainly performance. We don't want to negatively 21 impact that. 22 Certainly, you don't want to have a whole 249 1 lot of added time, just because of this added 2 functionality. 3 Also, I think that a thing that should be 4 considered, is the whole popularity of the solution, 5 meaning that we think about it again, take another 6 look at what we want something to look like. 7 What are some of those features that we 8 want up front? For instance, we're saying fruits 9 and vegetables, and we want the system to be able to 10 change to accommodate, maybe, something else. 11 These are some of the things we need to 12 be thinking of. Ideally, we should brainstorm, put 13 everything on the table, and then go for it, versus 14 not really thinking it out, having to go to the pilot 15 and maybe having to go to the pilot again. 16 Those are just a couple of things that 17 come to mind. 18 MR. KLEIN: Again, I think you go back to 19 cost. I think these gentlemen can probably build you 20 anything you want. 21 Inner city stores are operating with 22 state equipment. That equipment is giving you a 250 1 dollar value, only. 2 You have no idea what that recipient 3 bought. There is no platform. They're not working 4 with a third party processor. 5 They're just working through the state 6 equipment. So, if you get a dollar number, if it's 7 acceptable to you to know that Recipient A went 8 through and spent $25 and maybe they can just give 9 you a simple key that says, out of that $25, $5 was 10 a healthy choice, but I don't have -- I don't know 11 how much it costs for them to start telling you, by 12 the way, that $5 represents broccoli, asparagus, and 13 brussels sprouts, because that is not in the system 14 today. 15 In New York City, you do have the 16 majority of your retailers on state-operated 17 equipment, which is not operating through a third 18 party. 19 MR. RELICH: Can I just actually add to 20 that? Basically, that's the system you're talking 21 about building, where it would not track the type of 22 foods. Tracking the type of foods, that's why the 251 1 rules need to be very simple, so that an idiot could 2 understand what's a healthy food and what isn't. 3 Otherwise, it will be very complex. 4 MS. WALTERS: Okay, in looking at a 5 healthy incentive pilot, here are some admin costs 6 that I see for the states: 7 First of all, there are six states that 8 pay a retailer fee. We all know that retailers 9 incur transactions when they process an EBT card, so 10 six states reimburse the retailer, in part, for 11 every transaction. 12 So you put more money in the system and 13 that equals more transactions and increased costs 14 for the state and the Federal Government. The other 15 issue that I see from our perspective, is a training 16 issue. 17 Let's say we do modify the point-of-sale 18 equipment and the three small retailer farmers have 19 to press a button, $10, eligible food or healthy 20 food. What if I'm a farmer and I forget or I push 21 the wrong button? They're going to call and they 22 are going to say, oh, we should have done this. 252 1 So now we're looking at adjustments, 2 possibly. Who is going to make those adjustments? 3 Are the state staffs going to be involved? 4 Are we going to make a referral over to 5 our EBT contractor, to push that adjustment through? 6 Also, for training for farmers, like everybody else, 7 farmers are very busy, they're putting crops in the 8 ground, harvesting to earn money for their families. 9 The window of time for training, is 10 January through March. If I try to roll out a new 11 incentive program in July or August, they're not 12 going to want to hear from me. They don't want to 13 read anything from me during that time period. 14 It's going to be very crucial. I'm 15 trying to do training and explain this, and January 16 through March is my window of opportunity. 17 MR. COSTA: There's another cost in some 18 states. I don't have the exact number, but there 19 are several states that, in fact, reconcile 20 transaction data against the authorizations that 21 their states have put out. 22 So there could be a mismatch in the 253 1 authorization amount that was provided. In several 2 states, that means a modification to their systems, 3 at potential additional cost. 4 A potential scenario, is that there is an 5 additional cost that has to be reconciled out. That 6 would be an additional cost, also. 7 Getting down to some of the nuances of 8 what happens, for example, EBT allows for refunds. 9 If you develop a pilot, you want to develop a 10 transaction and when you have an accumulation of $10 11 in healthy food items, then you get a discount. 12 If I took home a bag of potatoes, but 13 when I got home, they were in excess of the 14 threshold and then I got back and found out that 15 half the bag was rotten, I came back, they don't 16 have an exact food item, they need a refund for me, 17 pushing it back under the threshold again. 18 So there are nuances that will have to be 19 thought out. They've got to look at anything that 20 we're trying to accomplish here and how complicated 21 we need to make this. 22 MR. WENNING: I would just add one thing, 254 1 that I think Jan made an important point about the 2 cost factor. The retailers will want to know, 3 particularly in those six states, and probably in 4 other states, what are going to be the cost impacts 5 on their operations. 6 I think you all know that there is cost 7 neutrality in the earlier EBT legislation. I 8 wouldn't be surprised if that became an issue, as 9 well. 10 The other issue that I talked a little 11 bit about earlier, was that fresh fruits and 12 vegetables and other healthy foods -- I think one of 13 the complicating factors that you're going to have to 14 take into consideration, is, there are a lot of 15 retailers out there who are developing wellness 16 programs within their supermarkets and they're 17 promoting healthy food products that go beyond fruits 18 and vegetables, and that there will be, I think, 19 whether it's defined as processed or unprocessed, 20 what's good processed, what's healthy, I think that 21 there will be an opportunity that they will want to 22 explore in having what is, quote, nutritionally 255 1 verified by their nutritionists as a list of eligible 2 healthy food items that could be used for this 3 program. 4 MR. BAKER: As you may have noticed, I 5 think it's a little early. As we start to pin down 6 what these things look like, you can get into these 7 details, the detailed level of what these factors 8 are, as talking about costs. 9 I'm not quite sure that this solution that 10 we're talking about, is a comparison to EBT 11 implementation. You don't have that available to 12 you. 13 What we're saying, is, we should expect 14 some economies of scale. 15 MR. CARLSON: In the conversation so far, 16 I've focused on the EBT card. Are there other 17 approaches that don't have to be on the EBT card, if 18 we had a separate debit card? What sort of historic 19 value that would have. 20 Would that make life simpler? Or is it 21 indifferent? 22 MR. BAKER: Since I have the mike, 256 1 realizing EBT takes me into electronics, but, 2 regarding another card, without diving into this or 3 having the opportunity to dive into it, it would just 4 make sense that you have a solution that has been 5 proven in testing, and, depending on the level of 6 complexity that we're looking for, it could be 7 enhanced to support this. 8 In my opinion, because of the similarities 9 between the two, one can be compared to the other. 10 We're talking about having one card versus two cards, 11 gets you a lot of benefits. 12 In addition to that, the cost associated 13 with that second card -- and I know we're probably 14 dicing and slicing different ways -- that's 15 certainly one way you can look at it. 16 MR. ANGEROSA: I agree with Errol on the 17 one card, with one major difference. When you talk 18 about a second card, a card that's going to support 19 nutritional purposes, instead of having a complete 20 optional program to try to incent people, they can 21 use the program or not, if they so desire. 22 You can implement whatever you please, 257 1 but if you have another card, you decide that 2 they're going to get this amount of money, and they 3 have to get nutritional food. It's a completely 4 different approach and much more rigid than putting 5 it on the EBT card, and saying, we've set up this 6 program, you're free to take advantage of it or not 7 and we can build in additional incentives. With one 8 card, you don't get that. 9 MR. CARLSON: Any other views on that? 10 (No response.) 11 MR. CARLSON: I have some colleagues who 12 must be borderline criminals. They come up with 13 some really interesting questions. It just never 14 occurred to me, but the question is, we're talking 15 about not unreasonable amounts of money. 16 Does that have the potential for new 17 threats to SNAP integrity? Are there ways to 18 identify what those threats might be, and mitigate 19 them? 20 MR. ALON: One of them is that the client 21 can use that to get the credit back. That's one way. 22 MR. CARLSON: That's exactly the sort of 258 1 example my colleagues raised. We create an 2 opportunity here. 3 Under the current rules, there is a fixed 4 amount, initially, provided each month. Is this kind 5 of incentive program, with the amount of money you 6 can get, might be dependent on the amount of 7 purchases or charges to be made, and there's a new 8 incentive for me to enter into collusion with you and 9 say, let's pretend that I bought $65 worth of healthy 10 food, and I'll split the difference. 11 MR. ALON: You can put a cap, to make it 12 more complicated, as part of the formula. Then you 13 have to -- the retailer -- I mean, right now, if the 14 retailer -- what are they going to do, redeem non- 15 healthy for healthy? 16 MR. KLEIN: I think the chance of fraud, 17 first, of all, is prevalent now. It is easy, if a 18 recipient wants to work with the cashier, in an 19 independent store, where there are very few systems 20 and controls. 21 We have actually documented stores that 22 have lost close to $200,000 in six weeks time, by 259 1 some very creative cashiers, who really know the 2 state system better than the state. They were 3 crediting and debiting funds within the hour before 4 the state swiped, to see what was actually there, 5 and were under everyone's radar, till we went back 6 and looked at it. 7 There are controls in your hands now, to 8 look at fraud. You know what an average transaction 9 is, you've documented over the years, what stores 10 normally do and whether their normal transaction is 11 $2.49 or $25.99. 12 A little more work on that end, would 13 certainly tighten some of that up. I don't think 14 you will have any more fraud than you have now, 15 because the fraud factor is still there. 16 But it is within the control of FNS to 17 monitor those transactions, but there has always 18 been a staffing issue of how many people can we have 19 to look at some of these things. But you do have the 20 control. 21 MR. CARLSON: It's interesting to observe 22 that we've probably got a decade's worth of 260 1 experience. This is a whole new wrinkle on that. 2 I'm not sure this happened. 3 MR. RELICH: To add to what Mitch is 4 saying, most of the fraud you have now, is 5 trafficking. Typically, it happens, but, very 6 rarely does it happen in the commercial world; we 7 rarely hear it at all, to be quite honest. 8 You will have people that will be 9 creative and take advantage of it, to a certain 10 extent. You can find people selling the benefits 11 for one reason or another, to get cash, but that's a 12 fact of life, I think. 13 MR. COSTA: It goes to how much incentive 14 you give them to create the fraud situations, a sort 15 of cap on the amount of healthy food items that you 16 purchase during the period. 17 One thing about being able to see that in 18 a pilot -- I heard timeframes this morning, for 19 pilots. I think that if you wanted to have your 20 pilot have some type of evaluation of the fraud 21 potential, that you have to have a long pilot, and a 22 12-week period of time would not be enough -- maybe, 261 1 if you're real creative. 2 (Laughter.) 3 MR. BAKER: I think a benefit that you 4 have, is piggybacking and leveraging the fraud 5 deterrent. Piggybacking on EBT, you have the 6 benefit of fraud tracking tools, so it's not like 7 you're going to have to recreate it. 8 What will have to happen, is, upgrade of 9 those tools, so that you try to stay ahead of those 10 out there that might want to commit fraud. 11 In my opinion, I think there should at 12 least be a small expectation that you might see a 13 slight increase in fraud, if there's a new feature 14 that you're adding. 15 So there is some risk there. It 16 shouldn't be too much. Again, it's not like you're 17 going to have a system in place. 18 MR. CARLSON: I have to confess that I'm 19 not sure I know what this next question means. 20 (Laughter.) 21 MR. CARLSON: But if you have an answer 22 to it, let me know. 262 1 The question is whether, in the kind of 2 solution that Dick was describing -- we know there 3 are alternative solutions. Is it reasonably 4 possible to identify the flows to the recipient, the 5 retailer, the banking system, the Federal Government 6 and recognize that this is separate from the regular 7 SNAP benefit, possibly accounted for, and it's going 8 to retain its identity, and is visible to all the 9 players? 10 Does that it sound like it, a) needs to 11 happen, and, b) can happen? 12 MR. RELICH: It's kind of interesting to 13 look at. It says "cost and benefits." Cost is 14 usually very easy to identify, particularly when 15 you're a retailer and you're selling groceries. You 16 know what your costs are. You know what you expect 17 to see. 18 From a banking aspect, that's easy to 19 see, too. Gene mentioned some issues, but all of 20 those are easily surmountable. 21 When you're talking about benefits, 22 that's kind of a loose term. Obviously, the client 263 1 has benefits. They're getting more for their SNAP 2 dollars. That's very valid. 3 The retailers obviously have value, 4 because they're selling more groceries and 5 everything else. 6 I'm sure the processors would like it, 7 because the return from the government is positive. 8 I'm sure the consultants would like it, because the 9 consultants helped build it. 10 (Laughter.) 11 MR. CARLSON: We all do what we do best. 12 MR. COSTA: I think it was pointed out 13 earlier that you can identify the dollar amount and 14 track it. What I think you may lose sight of, is, 15 when you get to the ten percent or 20 percent or 16 whatever you put the incentive level at, you want to 17 track how that is spent. 18 I think that's where you get into the 19 complexity on the system and how the next 20 transaction is debited against that very small 21 amount. So you're adding a level of sophistication 22 that the systems handle multiple types of benefits, 264 1 and, right now, states handle that, saying this 2 benefit occurred before this one, whatever the state 3 laws would be, as far as the order, but it will carry 4 a level of complexity that's not needed to track how 5 that $1.30 on that discount, was subsequently spent. 6 MR. RELICH: If you look at food stamp or 7 SNAP dollars, it's a single bucket. If you try to 8 track a specific amount, it's going to give you a 9 huge amount of complexity, and it's not going to 10 give you payback. I don't think you want to do 11 that. 12 MR. CARLSON: Let me turn it over to the 13 audience for you to ask all the questions you 14 couldn't ask or wanted me to ask. I see at least 15 one. 16 MS. SHAHIN: I don't know if I'm supposed 17 to be asking all of these, but I have them, anyway, 18 so I'm going to do it anyway. 19 The answer to this is, complexity costs 20 more money, but is there any reason why, to get to 21 the point you just made, Gene -- I bought $10 worth 22 of fruits and vegetables, and so I get my incentive, 265 1 which is 20 percent on that, so that's given to me as 2 a credit. 3 Those $2, I know A, B, C, can handle 4 various accounts in this, that, and the other, but 5 this would be the healthy food account, so $2 would 6 go in. That's what I can buy with that $2, is 7 healthy food. 8 So when I buy healthy food, I get dollars 9 to buy more healthy food, in the healthy food 10 account. I have a cash account, I have a food 11 account, I have my healthy food account. Could you 12 do that? 13 MR. COSTA: I'm going to do the very 14 sensible thing now and give this microphone to the 15 processor. 16 (Laughter.) 17 MR. ALON: Yes, it's possible. The 18 client will have $2 in one account and the $2 can be 19 used only against the healthy food items. Again, 20 this is more for the retailer. At the next shopping 21 time, if they want to use the healthy food account, 22 they will have to make sure that's in there. 266 1 MS. SHAHIN: This will probably get to 2 your point about delaying the gratification, being 3 problematic, as well. It may not be really 4 friendly, either, but I was actually thinking along 5 the lines of, if you've got $400 in your food 6 account, just keep on buying that healthy food out 7 of that food account, and that starts building that 8 healthy food account, so it gets bigger, and, by the 9 end of the month, when normally you weren't buying 10 food, you could buy some fresh fruits and vegetables 11 at the end of the month, because you would have 12 earned some money. 13 It's just a thought, and it sounds like 14 it would be complicated for the retailers, and I'm 15 not sure it would be all that attractive to the 16 clients. 17 MS. WALTERS: Jessica, I thought I would 18 add on to that one. Ideally, that would be great, 19 but how does little old Iowa, without projects, how 20 do we get enough leverage to get the third-party 21 processor to make a change, just for Iowa, for food 22 stamps? 267 1 You know, from the farmers market 2 perspective, it's a small percentage in the big 3 world. 4 MS. SHAHIN: You're absolutely right. 5 I'm looking at it much more from a pilot 6 perspective, than I am from the long-term, as to 7 what the pilot would be, and whether or not having 8 that additional money, made a difference. 9 In the end, it might not be a practical, 10 natural application, so, in that sense, it's 11 probably not a good idea. 12 MR. BAKER: I think it's a good idea. My 13 only thing with it, is, it could make it a little 14 complicated. 15 MS. SHAHIN: That was a very gracious way 16 of saying no. I like that. 17 (Laughter.) 18 MS. FOERSTER: I have two sort of 19 straight-out questions: If the WIC food package is 20 changed to having a certain dollar value for fruits 21 and vegetables that the state determines, the state 22 gets to say if it's fresh or canned or whatever. 268 1 Aren't all of your systems having to be 2 adapted now or in the next few months, to handle 3 this new benefit? So, is there an opportunity to 4 piggyback on the fact that the WIC food package is 5 doing something extremely similar? 6 Did you hear my question? 7 MR. KLEIN: I think you're also assuming 8 that your program is going to be identical to the 9 state's. The other problem is that you're going to 10 have 50 states with 50 different programs, so there 11 is going to be no consistency across the board. 12 On the WIC side, it is a check that comes 13 in, so the check comes in and there's not much to do 14 with your system, as long as the cashier understands 15 which item is approved by that state. 16 It's a matter of just having that cashier 17 training. Really, you don't have to do, on smaller 18 stores, any system change at all. 19 What I was getting at, is, the states 20 will each decide which items they are going to 21 reimburse with that $10, and so the back end is 22 already being done to handle that by different types 269 1 of stores, so some of that technology development or 2 software development, it seems like is going to be -- 3 it's not going to be identical. 4 That's going to set the limits on this 5 program. I'm just saying that you're going to have 6 some experience, and I wonder how that might set up 7 the work that we're talking about. 8 MR. RELICH: Is Aaron McBride still here? 9 (No response.) 10 MR. RELICH: What's happening -- and this 11 is actually a big issue now -- with WIC EBT, which 12 went on in a number of states -- when they looked at 13 fresh fruits and vegetables for cash value vouchers, 14 what they've been discussing, is, the retailers have 15 been adamant that this will be a common list. 16 They're not going to have one state 17 decide that apples, green apples are good, but the 18 other one decide they don't like green, and red 19 apples are good. 20 Basically, they're saying it has to be a 21 common list; there is just no way you can have the 22 variability. 270 1 One of the funniest things in WIC that 2 has come up, is, bananas can be substituted for baby 3 food in the WIC program. What's come is, is that a 4 fresh food or is it a vegetable or is it baby food, 5 and how do you handle it? 6 WIC EBT made the choice that it's not 7 baby food; it's fresh fruits and vegetables, and 8 they had to make those choices, from a systematic 9 standpoint. 10 You can't have: Monday we do this, 11 Tuesday, that. 12 MS. FOERSTER: Could I ask a second 13 question? 14 MR. RELICH: Well, make it simpler. 15 (Laughter.) 16 MS. FOERSTER: On farmers markets, I 17 would of have thought that if the benefit is being 18 delivered through EBT, that a farmer or a farmers 19 market, would just be able -- the patron would be 20 able to get that same benefit when she shops, 21 whether she's using farmers market bucks or whatever 22 she's using, that she would potentially get more put 271 1 back on the card for future purchases. 2 I'm not understanding why that would not 3 work for farmers markets. 4 MS. WALTERS: When one uses EBT benefits 5 at a farmers market, they're using that wireless 6 point of sale. On that wireless point of sale, you 7 have basically four tenders: EBT, debit, credit, or 8 cash. 9 I can swipe the card, hitting EBT, but I 10 have no way to differentiate whether or not or if 11 the customer is buying healthy food or something 12 else. There has to be some programming. 13 DR. PALMER: I had mentioned earlier that 14 I was concerned about fraud, if we went to any kind 15 of paper, because paper is sold; that's routine. I 16 know people do it. 17 There's still some fraud that goes on 18 with food stamps in small bodega-type stores, where 19 people have the need for beer or whatever, and they 20 know the person who's running it, but. But, you 21 know, it's relatively small and it's all client- 22 based. 272 1 But if I'm calculating this right -- and 2 I'd like you all to tell me -- won't fraud be even 3 more likely, because it would be motivated, because 4 there's a motivator now for the small business 5 owner, because if I charge all this as fruits and 6 vegetables, I still get some money that I would have 7 gotten, but now my consumer has all this extra money 8 that they also can spend with me. Am I correct in 9 presuming that there's even more motivation for fraud 10 that we'd have to be concerned about? 11 MR. RELICH: To be quite honest, I don't 12 think there's more motivation. If client can sell 13 the benefits, they're going to sell them. It takes 14 a retailer and it takes a client; it takes collusion 15 now. 16 DR. PALMER: What I'm saying, is, now the 17 retailer makes money off of it. 18 MR. RELICH: They've always made money. 19 When retailers do trafficking, they're paying 50 20 cents on the dollar, so if the client comes in, the 21 client, if he's smart, he's going to want 55 cents 22 now, because for a healthy item, you get another 273 1 five percent. 2 That retailer, if he's a bodega and has a 3 supply of fruits and vegetables and makes everything 4 a healthy item, what we've learned, is that these 5 people who do fraud, are very creative. They catch 6 on very quickly. 7 To be quite honest, you're not going to 8 have that much more. You'll have some, but it's 9 going to be rare, in relation to the overall 10 program. 11 If you look at it, fraud has gone down 12 quite a bit. On paper, it's gone down from four 13 percent to two percent per month. 14 MR. KLEIN: We've met with FNS over the 15 years, and, again, you have the best systems in 16 place. If I'm a bodega and you're looking at my 17 sales, and, all of a sudden, I have $25,000 a week 18 of produce, some bells should be going off 19 somewhere. 20 The fraud that exists today, is more 21 advantageous to the retailer. You're just 22 sweetening the pot a little. It's not that fraud 274 1 doesn't exist today. People sell the benefits, as 2 I've mentioned; people trade in the benefits. 3 If you go into bodegas, you'll find 4 stacks of cards that are held till the following 5 month, when the benefits are resumed. Everything, 6 though, is in place to monitor that. 7 WIC, on the other side, does a relatively 8 good job in undercover buys. It goes out and sends 9 people into stores. 10 Even in a bodega, a bodega that will risk 11 playing with food stamps, but they will not risk 12 playing with WIC. 13 Tim is still here. You run the risk of a 14 $40,000 fine, if you violate the rules and there's 15 any issue of fraud or miscalculation. 16 We have seen $40,000 violations for less 17 than $2. The point's been gotten across to small 18 stores. I don't think you'll have a bigger risk, 19 but, again, it is in your control to monitor, as you 20 look at the purchases. 21 As you mentioned before, you've got ten 22 years under your belt, of what stores have done, how 275 1 they have performed, and it's not hard to see. I 2 mean, you're not going to see a 25-cent fraudulent 3 transaction. That's probably an error on behalf of 4 the clerk. 5 But once people start to really get 6 involved in the fraud, they become really piggish 7 about it and you'll see those numbers ratchet up. 8 MS. WALTERS: In Iowa, the Iowa 9 Department of Agriculture, who administered the 10 farmers market nutrition program, they go out and 11 make routine checks on the vendors and the farmers 12 markets. Also for Iowa, for EBT wireless projects, 13 we have two of us and we go out and do random checks 14 throughout the week and on weekends. 15 In the four years that we've been doing 16 the program, I've had one phone call of a farmer 17 ringing himself up for another one, who wasn't an 18 EBT retailer. So we had a little sit-down talk, and 19 she promised me she would never do it again, and I 20 said, well, if you do, you won't be making any more 21 EBT. 22 So, we tried to go around and really try 276 1 and check on people, stand back and observe, before 2 going up and introducing ourselves. 3 MS. WOODFORD: Debbie Woodford, with the 4 WIC program. I just want to clarify the earlier 5 discussion about the fresh fruits and vegetables and 6 whether or not the decision between states, will 7 differ with regard to the fresh fruits and 8 vegetables. 9 There won't be any national, because the 10 only exclusion in the WIC, is for white potatoes, 11 but, on the other option, with regard to canned, 12 frozen, and dried, they will vary by state. That's 13 a state decision, as to whether or not they will 14 have them. 15 So I think the question was whether or 16 not they will be different lists, and there could 17 be, depending upon that state's decision with those 18 processed items. 19 But for fresh fruits and vegetables, 20 everyone -- there are no exclusions, other than 21 white potatoes. 22 MR. CARLSON: Any other questions? 277 1 (No response.) 2 MR. CARLSON: Any last words from the 3 panel? 4 (No response.) 5 MR. CARLSON: If not, let me again, 6 extend my thanks to all of you. 7 (Applause.) 8 MR. CARLSON: We want to do only two 9 things, as we wrap up: First, we want to extend my 10 thanks to all of you for sticking it out, for 11 actually making this an interesting day. 12 I really appreciated the participation of 13 all of you. Thought-provoking questions were 14 raised, as well as all of the comments that came 15 from the panel itself. 16 The next steps, as I have indicated on a 17 couple of occasions, are a little murky. I think, in 18 our highest expectations, we would have walked out of 19 here this afternoon with answers to all the questions 20 we posed. I think we can walk out with questions to 21 all the questions we've posed. 22 But we'll do our part in processing this, 278 1 and we'll work towards more public release of our 2 conclusions and in providing opportunities to jump in 3 and participate in a pilot. 4 The last thing I want to do, is turn the 5 microphone back to Jessica and allow her to speak. 6 MS. SHAHIN: I just wanted to add my 7 voice and say thank you to all the panelists, also, 8 to all the people in the audience that participated. 9 The questions that you asked, I'm personally a little 10 overwhelmed. 11 I, like Steven, had kind of thought maybe 12 we'd come out thinking, oh, okay, we know what we can 13 do; we've got our path. But you guys asked us 14 questions and that wasn't the way it was supposed to 15 work. 16 But I think what this told us, was that 17 we've got quite a job ahead of us, but it's still 18 exciting; it's just as exciting as it was this 19 morning, maybe even more so. 20 I also wanted to thank, specifically, a 21 lot of folks who made clear that this is an 22 important program, that does make a difference in 279 1 people's lives. I appreciated the way that many of 2 you actually spoke to people who receive these 3 benefits, thought of them as a part of this process, 4 thought about the communities, the client 5 communities, the retailer community, the industry 6 community, the farmers markets and the states. 7 That's when I think you start to 8 understand how really big and complex this is, 9 because there are so many stakeholders involved. In 10 the end, if we keep our eye on the client that we're 11 serving, we're going to get what we need out of this. 12 That's going to be the key today. 13 I think that some of the observations 14 that have been made, have been really maybe 15 surprising to some of us. That would be me. Maybe 16 not so surprising to others, but I just did want to 17 take a moment to say thank you so much. It's been a 18 long day, a full day, it provided us a great deal of 19 information. 20 Thank you very, very much for taking this 21 time, especially you folks who have come again. 22 (Applause.) 280 1 (Whereupon, at 4:10 p.m., the symposium 2 was concluded.) 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22