Part 1 |
Save page Remove page | Previous | 1 of 2 | Next |
|
|
small (250x250 max)
medium (500x500 max)
Large
Extra Large
Full Size
Full Resolution
|
This page
All
|
COMPLETED United Stales Department of Agriculture Food and NubWon Service Office of Analysis and Evaluation M7V-A-M Aft.2:C73 Commodity Letter of Credit Modification Demonstration Evaluation Final Report tf T ABEL, DAFT & EARLEY 1410 KING STREET. ALEXANDRIA, VA 22314 TEL: (703) 739-9090 FAX: (703) 739-9098 COMMODITY LETTER OF CREDIT MODIFICATION DEMONSTRATION EVALUATION FINAL REPORT by Lynn M. Daft and Donald W. Westfall Abel, Daft & Earley Alexandria, Virginia submitted to: John R. Endahl Food and Nutrition Service U.S. Department of Agriculture 3101 Park Center Drive, Room 204 Alexandria, Virginia 22302 March 1992 T EXECUTIVE SUMMARY Introduction The National School Lunch Program (NSLP) provides Federal assistance to the feeding programs of the nation's elementary and secondary schools. This assistance is provided in two forms: cash and donated commodities. For the past ten years, beginning in School Year (SY) 1982/83, 25 school districts throughout the nation have participated in a demonstration of an alternative to the existing commodity donation component of the NSLP. Under this program, school districts are authorized through letters of credit to make their own purchases of specified foods in place of receiving donated commodities purchased by USDA. The commodity letters of credit (CLOCs) are used to purchase foods from local commercial sources. In response to criticisms of some elements of CLOC, the program was modified in SY 1990/91. This report describes the results of an evaluation of the effects of these modifications. The evaluation began in September 1990 and was conducted for the Food and Nutrition Service of the U.S. Department of Agriculture by Abel, Daft & Earley of Alexandria, Virginia. Background The CLOC alternative arose out of a desire by some School Food Authorities (SFA's) for changes in certain features of the commodity donation program as it was operated in the 1970's. For example, some commodities were received in forms that could not be conveniently used in preparing meals (e.g., whole turkeys, large blocks of frozen ground beef, and 100 pound bags of flour). Some of the foods were of low acceptability to students and therefore difficult to use. Some schools lacked control over the size and timing of commodity deliveries. As a result, donated commodities often had to be stored for as long as several months. For these and other reasons, there had been a continuing interest on the part of some schools in alternatives to commodity donation. This interest ultimately led to a Congressional directive, as part of the 1981 Agricultural Appropriations Act, to conduct a 3-year study of two alternatives to the existing commodity donation program -- an all cash approach and CLOC. The study got underway in SY 1982/83 with 96 school systems from 29 States assigned to one of the three options under study: commodity donation, cash, and CLOC. The study ran through SY 1984/85. The cash school systems in the study were allocated cash equivalent in amount to what the USDA would have used in buying commodities for donation to them. The SFA's, were free to use this cash in support of their school feeding programs in any way they wanted. II CLOC SFA's also received cash equivalent to the value of the donated commodities that they would have received, but the expenditure of this cash was restricted to foods containing those commodities that USDA was buying for donation through the NSLP. In this way, CLOC funds were used to provide market support for the same commodities USDA was supporting through its purchases. CLOC funds were allocated among commodities in the same general proportion as USDA purchases. Authorization for the use of funds (through the issuance of CLOC's) was timed to coincide with USDA's purchases, as were deliveries of these purchases. To provide SFA's with maximum flexibility in their use of these funds, partial credit was granted for the purchase of a wide range of foods containing the particular commodity. The amount of credit was dependent on the share of product weight accounted for by the commodity. Final results of the evaluation of this initial demonstration of CLOC were published in 1986.-^ On the one hand, they demonstrated that CLOC was administratively feasible at the local level and that most of the SFA's participating in CLOC were pleased over its effects on their individual feeding programs. On the other hand, CLOC schools removed less of some commodities (e.g. beef, chicken, pears, and sweet potatoes) from the market. Related to this, CLOC schools bought more highly processed foods, in effect spending a larger share of their Federal assistance on processing and a smaller share on commodities. Results of the cash option were similar to those for CLOC, though the reduced level of commodity support was even more pronounced, at least for certain commodities. At the conclusion of this evaluation, the USDA recommended that CLOC be discontinued. However, SFA participants in CLOC successfully appealed to the Congress for its continuation as a pilot project. The program has continued to operate on this basis ever since. The commodity donation program has been improved in a number of respects over the past 5 to 10 years. This includes reduction in the fat and salt content of foods distributed , improved labeling and product identification, inc.eased product varieties, improved scheduling of deliveries, improved packaging, the adoption of unitized deliveries, and greater attention to nutrient analysis. In 1989, the Conference Report accompanying the Child Nutrition and WIC Amendments Act authorized the Secretary of Agriculture to design and test modifications of CLOC that would address shortcomings of the program identified in the earlier evaluation. Later in the year, the Secretary directed the Food and Nutrition Service to develop modifications of CLOC based on consultation with the affected interests and to conduct an evaluation of the effects of these modifications. * St. Pierre, Robert G., et. al. Evaluation of Alternatives to Commodity Donation National School Lunch Program: Final Report. Volume o, 1986. u Ml Commodity Letter of Credit Modifications The CLOC program was modified in a number of important respects, effective at the beginning of SY 1990/91. The program was modified for all 25 school systems that were still participating in CLOC. The most important of these changes were the following: • Specifying the commodity to be purchased. The USDA now determines on a case-by-case basis whether CLOC's are to be issued for generic commodities (e.g. "pears"), as they were in the past, or for more specific commodities (e.g. "fresh winter pears"). • Full and partial credit. Authorizations for the use of CLOC funds are now based on achieving a farm-equivalent weight M of the targeted commodity per dollar spent comparable to that of USDA purchases. This has the effect of narrowing the range of products for which credit is authorized and reducing the amount of partial credit allowed for many foods. • Timing of purchases and deliveries. The time periods allowed for CLOC purchases and deliveries have generally been accelerated and shortened to better match the schedule of the USDA commodity program. For CLOC purchases of fruits and vegetables the reduction was especially large. • Trades, rollovers, and substitutions. The rollover of small balances remaining in CLOC accounts and limited substitution among commodities, both of which were permitted under the original CLOC concept, have been discontinued. And, the trading of letters of credit of one commodity for those of another commodity, among participating CLOC SFA's are no longer required to be of equal dollar amounts. • Domestic origin. SFA's and their distributors are required to certify that all foods purchased with CLOC funds are of domestic origin. Evaluation Objectives and Methodology There are two central issues to be addressed in this evaluation, both related to the CLOC modifications adopted in SY 1990/91. They are: (1) The effect of these modifications on the performance of the CLOC program relative to the performance of the existing Commodity Donation Program. y Farm - equivalent weight is the weight of a raw agricultural commodity at the farm gate that is required to produce a given unit of a food product containing that commodity. For example, it takes approximately 1.25 pounds of raw apples coming out of the orchard to produce 1 pound of canned applesauce. • ■ • m Y (2) The effect of these modifications on the performance of the CLOC program now relative to its performance prior to the modifications. The Congressional authorization for these modifications made clear that it was looking for modifications that would overcome criticisms of the earlier program without undermining the CLOC concept's central strengths. These general issues were first translated into a list of 14 research objectives and then into a list of about 20 researchable questions. The list includes such questions as: Do CLOC and Commodity SFA's remove the same weight of each commodity per lunch served? Do CLOC SFA's buy more highly processed foods? To what extent are CLOC's not used or traded and how does this compare to the past? How do the level of benefits received by CLOC and Commodity SFA's compare? How does the timing of CLOC issuances, purchases, and deliveries compare with the timing of USDA purchases and deliveries? To what extent are CLOC's used to purchase foods that are not of domestic origin? What effect did the CLOC modifications have on labor use and costs and or rates of participation in the NSLP? How do CLOC SFA managers view the modifications? Of the 32 CLOC SFA's that participated in the original demonstration, 25 remain in the program. Because participation in the CLOC program is voluntary, those SFA's that remain in it apparently prefer CLOC to the existing commodity donation program. Furthermore, having participated in CLOC for nearly 10 years, these SFA's have had time to learn CLOC procedures and were therefore reasonably well prepared to adjust to the modifications. For purposes of this evaluation, 2Z additional SFA's were selected as matched control sites. They were matched to the CLOC SFA's on the basis of a number of criteria including size of enrollment, number of schools, percent nonwhite, and percent below the poverty line. With one exception, control sites were matched with CLOC sites from the same State. The one exception was made necessary by the large enrollment of the SFA against which it was matched. Each of 50 SFA's participating in the evaluation provided several types of data, some of it in highly detailed form. To determine the volume and value of food acquisitions and IV < use, each SFA provided records describing food purchases, commodity donation deliveries, and beginning and ending inventories for all foods containing any of 13 commodities for SY 1990/91. y The evaluation was restricted to 13 commodities to help make data collection more manageable and less costly. For the most part, this information was provided in the form of copies of invoices, delivery sups for donated commodities, and ending inventories. These data were periodically mailed to a data collection office where they were transcribed, entered into micro computers, and edited. Participating SFA's also provided operating data for their operations for SY's 1989/90 and 1990/91 through two mail surveys. A 100 percent response rate was achieved for both surveys. Information on the State commodity distribution systems was collected through a mail survey of the 50 State Distributing Agencies (SDA's) that distribute commodities through NSLP. All 50 agencies provided the information that was requested. Finally, limited information on the domestic content of foods purchased with CLOC funds was collected on-site by FNS Regional Office Staff from 18 CLOC SFA's. Most of the research questions have been analyzed through a comparison of mean values. For some questions these are comparisons of modified CLOC and Commodity SFA's; for other questions they are comparisons of CLOC SFA's before and after the modifications were adopted. Statistically significant differences have been determined using the two-sample t-test and the paired t-test, as appropriate. In answering some research questions, values have been adjusted through use of a regression model designed to eliminate the influence of a variety of exogenous factors such as enrollment size and wage rates. Evaluation Findings The most significant findings of the evaluation are those related to the two central issues described earlier, i.e., how have the CLOC modifications affected (a) CLOC performance relative to Commodity performance and (b) the advantages of CLOC as it has operated in the past? Overall, the findings indicate that a number of the major differences in the performance of CLOC and Commodity SFA's found in the previous study ai.d that were the focus of the CLOC modifications have been narrowed or eliminated. Furthermore, the principal attractions of CLOC were found to have been largely unaffected by these modifications. The more detailed findings are described below. Market removals. Almost no significant difference was observed in the farm-equivalent weight of commodity acquisition or use adjusted by the number of lunch V The 13 commodities are: beef, pork, chicken, turkey, apples, peaches, apricots, pears, green beans, green peas, potatoes, sweet corn, and tomatoes. VI T equivalents U (LEQ's), between CLOC and Commodity SFA's. In the case of only one commodity, pears, was it found that CLOC SFA's used a larger volume per LEQ than Commodity SFA's. Of all the detailed findings, this is perhaps the most central. It says that the modified CLOC program provides support for commodity markets over the period of a school year comparable to that provided by the commodity donation program. Product composition. Most of the differences in the product composition of acquisitions by CLOC and Commodity SFA's are related to the form in which commodities are donated. This indicates that when SFA's are given discretion in their choice of product forms, they frequently opt for a form other than that in which commodities are donated. Whole turkey, cut-up chicken, frozen peaches, frozen ground beef, and bulk apple juice are examples of products that accounted for a significantly larger share of Commodity SFA purchases. While there is some evidence that CLOC SFA's acquire more highly processed foods, the relationship is weak. Only in the case of chicken were Commodity SFA's found to have acquired products with a higher share of their cost attributable to farm value. Timing of purchase and delivery. Modifications in CLOC resulted in a substantial reduction in the length of the purchase period, especially for fruit and vegetable CLOC's. The average purchase period was reduced from 172 days in SY 1939/90 to 39 days in SY 1990/91. Nearly all CLOC purchases were made within the prescribed time limits. As a result, in SY 1990/91, CLOC purchases coincided very closely with the timing of USDA purchases. Although CLOC procedures allow for a delivery period of at least 3 or 4 months, most CLOC SFA's arrange for delivery within the purchase period to minimize the paperwork involved. This has resulted in CLOC SFA's taking delivery of their purchases 3 or 4 months before the corresponding commodity donations reach Commodity SFA's. Benefit levels. The overall value of the benefits received by CLOC and Commodity SFA's, adjusted for the number of lunches served, were found to be roughly comparable. For individual commodities, however, relatively large variations were observed. For example, Commodity SFA's received over 30 percent more pork than CLOC SFA's on a dollars per reimbursable meal basis. Most of these differences are due to operational differences fundamental to the two approaches rather than to any decisions made at the SFA level. A comparison of the value of commodity donation receipts by Commodity SFA's with an estimate of the value of individual commodities to which they were entitled reveaied y The lunch equivalent measure converts numbers of lunches and breakfasts and the dollar value of other food sales (e.g. a la carte and adult meals) to a common unit. It is based on the number of lunches that could have been produced for the same cost incurred in producing a combination of lunches, breakfasts, and other food sales. vi VH a relatively large variation in the share of entitlement received. This variation occurs both among commodities and among SFA's. Of the 25 Commodity SFA's, over one-third received commodities that deviated from their entitlement value by 20 percent or more, some higher and some lower. These variations in entitlement benefits received by Commodity SFA's result in part from SFA's refusing all or part of a particular commodity offered them, SFA's accepting other SFA's refused commodities, limitations on case-splitting for small SFA's, and variations in the methods SDA's use to calculate total entitlement. The implication of this finding is that there is a far larger degree of variation in the level and composition of commodity donation benefits than in CLOC benefits. CLOC utilization. Of the total value of CLOC's issued in SY's 1989/90 and 1990/91,99.8 percent was spent. The CLOC modification had no appreciable effect on this share; in fact, the share spent increased marginally in SY 1990/91. CLOC trades among participating SFA's have not been used extensively, though the level of activity increased between SY 1989/90 and SY 1990/91 as the share of CLOC value traded rose from 0.2 percent to 0.4 percent. CLOC refusals and rollovers, option.1; that were discontinued under the modifications, made it possible for CLOC SFA's to substitute among commodities and to more completely exhaust CLOC funds in SY 1989/90, though neither option was extensively used. Labor use. No significant differences between CLOC and Commodity SFA's were observed regarding changes in either labor hours or labor cost. This is an indication that the CLOC modifications did not adversely affect labor use. Both measures were adjusted by the number of lunch equivalents. Participation rates. The CLOC modifications had no significant effect on rates of participation in school lunch, school breakfast, or a la carte sales. Effect of CLOC on other commodity donation programs. Of the 50 State Distributing Agencies (SDA's) that are responsible for the administration of commodity donations through the NSLP, ten would have to make major changes in their distribution systems. On average, these SDA's devote approximately 15 person-years to warehousing and delivery in their present systems. Changes in program administration would be required for all SDA's, though this was not a topic of study in this evaluation. The administrative feasibility at the state level and its impact on other commodity donation program remain major unanswered questions. CLOC SFA manager views. Food service managers of CLOC SFA's reported that the CLOC modifications had comparatively little adverse effect on their operations. Increased administrative costs associated with dealing with changes in the use of full and partial credit was the most frequent complaint followed by some dissatisfaction over the purchase period for some commodities and over the discontinued use of refusals and rollovers. Dissatisfaction with the purchase period appears to be associated with those purchases that were required to be made very early in the year, before school started. Domestic origin. An on-site survey of CLOC purchased inventories held by 19 CLOC SFA's in the spring of 1991 revealed that the labels on a significant portion of two of the foods surveyed - canned tuna and apple juice -- indicated that they contained vii nn imported product. This occurred despite SFA attempts to require distributors to provide only domestically produced foods. It is concluded that: (a) the CLOC approach cannot ensure that foods purchased are entirely of domestic origin and (b) that the incidence of CLOC purchases of foods that are not of domestic origin is highest when USDA purchases foods that are difficult to buy through commercial vendors or are significantly more expensive than comparable foods of foreign origin. Study Limitations Sample Size and Representativeness. The size of the evaluation sample -- 25 CLOC systems and 25 Commodity systems - is small relative to the size of the universe of over 15,000 SFA's. While this sample includes districts from various parts of the country with varying enrollment sizes in both rural and urban settings, it was not designed to be nationally representative. A limited comparison of sample district characteristics with national totals indicates that the sample over-represents larger school districts and those that participate in the School Breakfast Program. However, the CLOC sites and comparison sites of Commodity SFA's are very closely matched to each. Commodities Studied. The scope of the agricultural impact analyses was limited to 13 major commodities. While these commodities accounted for approximately 75 percent of the total value of entitlement commodities, restricting the evaluation in this way may have overlooked potential market impacts for smaller purchase programs. Administrative Feasibility. Questions on the administrative feasibility of the CLOC system were limited to the ability of CLOC SFA's to adapt to the CLOC modifications and to the effects of the modifications on local program administration. Issues related to administration of the CLOC program at the State or National levels were not addressed. viii X CONTENTS Page Executive Summary i List of Tables xii List of Figures xvii Acknowledgements xviii I. Introduction 1-1 A. History of the Commodity Donation Program 1-1 B. Program Operations 1-3 C. Results of Past Evaluation MO D. Modifications of the Commodity Letter of Credit 1-13 II. Objectives of the Evaluation II-l A. Issues to be Examined II-l B. Research Questions II-2 III. Evaluation Methodology III-l A. Evaluation Sample III-l B. Data Collection III-2 C. Analytic Techniques III-8 D. Special Measurement Issues and Problems Ill-12 IV. Participating School Food Authorities IV-1 V. Evaluation Findings V-l Questions: 1 Weight of commodity removed per KLEQ -- CLOC/Commodity compared V-l 2 Value of mix of products removed from the market -- CLOC/Commodity compared V-5 3 Share of value of foods attributable to raw commodity value -- CLOC/Commodity compared V-19 4 Share of value of CLOC's net spent V-22 ix X Page 5 Share of entitlement received by Commodity SFA's .... V-28 6 CLOC issuance/commodity donation compared per reimbursable meal V-38 7 Weight of commodity per dollar of CLOC issued « 1989/90 and 1990/91 compared V-41 8 Share of value of CLOC issuance not spent - 1989/90 and 1990/91 compared V-45 9 Share of CLOC's refused, rolled over, traded, or not spent in 1989/90 compared to share traded or not spent in 1990/91 V-46 10 Timing of CLOC and USDA purchases compared ..... V-52 11 Timing of CLOC deliveries and donated commodity deliveries compared V-53 12 Pattern of CLOC purchases by month following date of issuance -- 1989/90 and 1990/91 compared V-60 13 Changes in labor cost, labor hours, and food costs per KLEQ -- CLOC/Commodity in 1989/90 and 1990/91 compared V-64 14 Rates of labor use by CLOC SFA's over time -- 1983/84-1984/85 and 1990/91 compared V-66 15 Changes in participation rates ~ CLOC/Commodity in 1989/90 and 1990/91 compared V-71 16 CLOC manager's views of program modifications V-75 17 Effect on other programs of nationwide adoption of CLOC V-77 18 USDA commodity purchasing criteria and practices .... V-86 19 USDA and commercial prices compared V-94 20 Domestic origin of CLOC purchases V-96 y Page 21 Effect of CLOC modifications on program administration V-103 VI. Study Limitations VI-1 A. Sample VI-1 B. Data VI-1 C. Methodology VI-3 VII. Conclusions VIM Appendix A: Participating School Food Authority Descriptive Data A-l Appendix B: Comparison of Differences in Products Acquired and Used by CLOC and Commodity SFA's, SY 1990/91 B-l xi LIST OF TABLES Page I. Introduction Table 1-1 Federal Support for School Food Programs, FY 1969 - FY 1990 1-4 Table 1-2 Food Donated to School Districts Through The Commodity Distribution Program, School Year 1990/91 1-6 II. Objectives of the Evaluation Table II-1 Objectives and Research Questions II-2 m. Evaluation Methodology Table III-l School Food Authorities Participating in the Commodity Letter of Credit Modification Evaluation III-3 Table III-2 Correlation Coefficients of Variables Used in the Regression Model III-ll IV. Participating School Food Authorities Table IV-1 Summary of Program Activity: 1989/90 IV-2 Table IV-2 Summary of Program Activity: 1990/91 IV-3 Table IV-3 Comparison of Participating SFA's on Selected Variables By Demonstration Alternative With National Totals IV-4 Table IV-4 Average Price Charged for Full Price Meal and Change in Average Meal Prices, 1989/90 to 1990/91 IV-8 xu Xlll Page V. Evaluation Findings Table V-l Comparison of the Mean Farm-Equivalent Weight of Commodities Acquired per KLEQ by CLOC and Commodity SFA's, SY 1990/91 V-3 Table V-2 Comparison of the Mean Farm-Equivalent Weight of Commodities Used per KLEQ by CLOC and Commodity SFA's, SY 1990/91 V-4 Table V-3 Comparison of Differences in the Regression Adjusted Mean Farm-Equivalent Weight of Commodities Acquired and Used per KLEQ by CLOC and Commodity SFA's SY 1990/91 V-6 Table V-4 Comparison of Differences in the Value of Products Acquired as a Percent of the Total Value within Each Commodity by CLOC and Commodity SFA's, SY 1990/91 V-12 Table V-5 Comparison of Differences in the Mean Percent Value of Acquired Products Derived From Each Commodity Attributable to Raw Commodity Value for CLOC and Commodity SFA's, SY 1990/91 V-20 Table V-6 Comparison of Differences in the Mean Percent of Value of Used Products Derived From Each Commodity Attributable to Raw Commodity Value for CLOC and Commodity SFA's, SY 1990/91 V-21 Table V-7 Comparison of Differences in the Regression Adjusted Mean Percent Value of Products Derived From Each Commodity Attributable to Raw Commodity Value for CLOC and Commodity SFA's, SY 1990/91 V-23 Table V-8 Mean Percent of the Total Value of CLOC Issuances in SY 1990-91 Not Spent V-29 Table V-9 Comparison of the Distribution of USDA's Reported Purchases, Deliveries to SDA's, CLOC Issuances, and Donated Commodity Deliveries to Study Sites, 13 Selected Commodities, SY 1990/91 .. V-33 Table V-10 Proportion of Total Entitlement Allocated to Selected Commodities for CLOC SFA's, SY 1990/91 V-34 xiii XW Page Table V-ll Mean Percent of the Value of Commodity Entitlement Received by Commodity SFA's, SY 1990-91 V-35 Table V-12 Comparison of the Mean Value of CLOC Issuances per KRM and the Mean Value of Commodity Donations per KRM, by Commodity, SY 1990/91 ... V-40 Table V-13 Comparison of the Mean Value of CLOC Issuances per KRM and the Mean Value of Commodity Donations per KRM, by Commodity, SY 1990/91 ... V-42 Table V-14 Comparison of the Farm Equivalent Weight of Commodity Removed from the Market per Dollar of CLOC Issuance in SY 1989-90 and SY 1990-91 V-44 Table V-15 Comparison of the Average Percent of CLOC Issuance Not Spent in SY 1989/90 and in SY 1990/91 V-47 Table V-16 Comparison of the Mean Percent of the Value of CLOC's Not Used in SY's 1989/90 and 1990/91 .. V-50 Table V-17 Comparison of the Date of CLOC Issuance and Expenditure with the Date of USDA Purchase, SY 1990-91 V-54 Table V-18 Mean Share of Donation Deliveries to SFA's by Month, by Commodity, SY 1990/91 V-57 Table V-19 Comparison of the Date of CLOC Issuances and Expenditure, SY 1989-90 V-61 Table V-20 Comparisons of Mean Changes in SFA Labor Hours, Labor Costs and Food Costs Per KLEQ for CLOC and Commodity SFA's, SY 1989/90 and SY 1990/91 V-65 Table V-21 Comparison of Differences in the Regression Adjusted Mean Changes in SFA Labor Hours, Labor Costs and Food Costs per KLEQ Between SY 1989/90 and SY 1990/91 V-66 Table V-22 Comparison of the Number of LEQ Per Labor Hour for CLOC SFA's in SY's 1983/84 - 1984/85 and SY 1990/91 V-70 xiv vv Page Table V-23 Comparison of Rates of Participation and Changes in the Rates of Participation In CLOC and Commodity School Feeding Programs Between SY 1989/90 and SY 1990/91 V-72 Table V-24 Comparison of Regression Adjusted Changes in the Rate of Participation in CLOC and Commodity SFA's between SY 1989/90 and SY 1990/91 V-73 Table V-25 Summary of CLOC SFA Managers's Opinions About the Impact of CLOC Modifications V-75 Table V-26 CLOC SFA Manager Opinions Regarding the Effect of CLOC Modifications on Operations, by Commodity, SY 1990/91 V-77 Table V-27 Value of Domestic Food Donations, by Program, 1985/86-1990/91 V-80 Table V-28 Commodity Donation Program Responsibilities of State Agencies Administering Commodity Donations Through School Lunch Program V-81 Table V-29 Number of SDA's and Share of Shipment by Type of Distribution System for NSLP SDA's, 1985/86 and 1989/90 V-83 Table V-30 Farm Gate Prices: 1989-90 and 1990-91 Marketing Years V-88 Table V-31 Volume and Value of USDA Commodity Donation Deliveries through the NSLP, SY's 1989/90 and 1990/91 V-90 Table V-32 Comparison of USDA Purchases Prices for Selected Products, 1990/91 V-93 Table V-33 Comparison of U.S. Domestic Production and Imports of Selected Food Items, CY 1988 and CY 1989 V-96 Table V-34 Results of On-Site SFA Visits to Identify Foreign Origin Foods V-99 xv y\ji Page Table V-35 Allocation of Staff Time in Administering the CLOC Program, SY 1989/90 and SY 1990/91 V-101 Appendix A. SFA Descriptive Data Table A-l Size Distribution of SFA's Participating in the Evaluation Compared to Size Distribution Nationally in SY 1988/89 A-l Table A-2 Food Service Program Responsibility 1989-90 School Year A-2 Table A-3 Food Service Program Responsibility 1990-91 School Year A-3 Table A-4 Number and Mean Share of Schools by Type 1989-90 School Year A-4 Table A-5 Number and Mean Share of Schools by Type 1990-91 School Year A-5 Table A-6 Participation Measures 1989-90 School Year A-6 Table A-7 Participation Measures 1990-91 School Year A-7 Table A-8 Reimbursable Lunches and Breakfasts 1989-90 School Year A-8 Table A-9 Reimbursable Lunches and Breakfasts 1990-91 School Year A-9 Table A-10 Food Service Options 1989-90 School Year A-10 Table A-11 Food Service Options 1990-91 School Year A-11 Table A-12 Kitchen Configurations 1989-90 School Year A-12 Table A-13 Kitchen Configurations 1990-91 School Year A-13 Table A-14 Average Meal Prices 1989-90 School Year A-14 Table A-15 Average Meal Prices 1990-91 School Year A-15 xvi yvi Page Table A-16 Federal Lunch Reimbursement 1989-90 and 1990-91 School Year A-16 Table A-17 Staffing, School Year 1989-90 A-17 Table A-18 Staffing, School Year 1990-91 A-18 Table A-19 Sources of Income 1989-90 School Year A-19 Table A-20 Sources of Income 1990-91 School Year A-20 Table A-21 Food Service Expenses 1989/90 School Year A-21 Table A-22 Food Service Expenses 1990/91 School Year A-22 LIST OF FIGURES Figure V-l Distribution of SFA's by Percent of Entitlement Received V-37 Figure V-2 Overall Comparison of CLOC and Commodity Delivery Patterns, SY 1990/91 V-58 Figure V-3 Apple Product Deliveries V-59 Figure V-4 Green Bean Product Deliveries V-59 xvu y\/u\ ACKNOWLEDGEMENTS A number of people contributed to the preparation of this report. Without attributing to them any responsibility for errors or omissions, the authors would like to thank one and all for their help. A very special thanks is due the food service directors and staffs of the 50 school systems that participated in the evaluation by supplying reams of information throughout the year. Without their dedication to improvements in school food service, as evidenced by their willingness to voluntarily take part in this demonstration, this evaluation wouldn't have been possible. Two other individuals who deserve special thanks for the contributions they made throughout the evaluation are Cherie Root and Ann Hallawell of Virginia Polytechnic Institute and State University. Having administered the CLOC pilot program for the past 10 years, the rich and unique experience they brought to the evaluation was of great value. Beyond supervising the collection and transcription of the thousands of food acquisition and inventory records used in the analysis, they served as general consultants throughout the project. The evaluation also benefitted from the periodic advice of Michael J. Puma of Abt Associates, Dorothy V. Pannell of Fairfax County (Virginia) Public Schools and Joanne L. Styer of Montgomery County (Maryland) Public Schools. Their advice, both practical and methodological, based on their extensive experience in school feeding and in the evaluation of school feeding programs was greatly appreciated. As in most studies, this evaluation has been constructed on a foundation of knowledge developed through earlier research. Two studies were particularly relevant and valuable in this regard. The initial evaluation of CLOC by Abt Associates, completed in 1986, provided a useful frame of reference, both for design of the evaluation and for comparison of results. V Robert G. St. Pierre and Frederic B. Glantz, two of the principals in the original Abt Associates evaluation, also made useful suggestions regarding the measurement of lunch equivalents. A second report of particular value in preparation of the data analysis and collection plan was prepared by Sheena McConnell and Harold Beebout of Mathematica Policy Research. Inc.2/ " St. Pierre, Robert G. et. al., Evaluation of Alternatives to Commodity Donation in the National School Lunch Program: Final Report. Abt Associates, Cambridge, Mass., prepared for the Food and Nutrition Service, USDA, 1986. V McConnell, Sheena and Beebout, Harold, Analysis Plan for the CLOC Modification Demonstration. Mathematica Policy Research, Inc.. Washington, D.C., prepared for the Food and Nutrition Service, USDA April 3, 1990. xviii KIY. We are indebted to those staff members of the Agriculture Marketing Service responsible for commodity procurement who took time to discuss their procedures with us. Our appreciation also to Alberta Frost, Director of the Food Distribution Division of FNS for her advice and help and to the FNS Regional Office Staff who collected information for analysis of the domestic content issue. Sheila Wilson of Abel, Daft & Earley maintained her cheerful composure throughout the typing of several drafts of this report, for which we express our appreciation. Last but certainly not least, John R. Endahl, the FNS Contracting Officer's Representative for the project, has our special thanks. Having participated in both the administration and the evaluation of the CLOC demonstration for several years, he brought to the project a balance of experience, insight, and analytic competence that was of substantial value. He provided critical advice and support throughout the project. xix yx L INTRODUCTION This report presents the findings of an extensive evaluation of modifications in the Commodity Letter of Credit (CLOC) program that were adopted by school districts taking part in the program in School Year (SY) 1990/91. For the past ten years, beginning in SY 1982/83, 25 school districts throughout the United States have participated in a demonstration ofCLOC as an alternative to the U.S. Department of Agriculture's (USDA's) Food Distribution Program to schools in the National School Lunch Program (NSLP). Under the program, school districts are authorized through letters of credit to make their own purchases of specified foods in place of receiving donated commodities purchased by USDA. Results of an earlier evaluation, ending in 1986, led the USDA to conclude that the CLOC approach did not warrant continuation. However, at the urging of school districts taking part in the demonstration and other associated interests, the Congress authorized the participating districts to continue in the CLOC program. On three other occasions since then the program has been extended and a measure is now pending in Congress to extend it once again. The issue of whether CLOC should be adopted on a broader scale or discontinued altogether remains unresolved. In the Conference Report language accompanying the Child Nutrition and WIC Amendments Act of 1989, the Congress authorized modifications in the CLOC program as a means of determining whether the most commonly voiced criticisms of the program could be overcome while, at the same time, retaining CLOC's principal advantages. The central purpose of this evaluation is to examine the extent to which these modifications have achieved that purpose. The report begins with a brief description of the background that led to the design of CLOC, its implementation on a demonstration basis, and results of the initial evaluation. Shortcomings of the program are discussed, as are the modifications that were adopted in 1990/91 to address these limitations. This is followed by a description of the methodology employed in conducting the study and a profile of the school systems that participated in it. Study findings are then reported for each of the research questions that are being examined. A final chapter presents major conclusions of the evaluation. A. History of the Commodity Donation Program Beginning in the 1930's when farm prices were severely depressed and the ranks of the unemployed were growing ever larger, the Federal government began donating surplus food to those who were poor and unemployed. Donation programs were viewed then, as they continue to be viewed today, as an opportunity to improve the nutritional status of people in need while at the same time strengthening the demand for commodities that are temporarily in surplus, and in that way supporting farm prices and farm income. Since their first use nearly 60 years ago, donated commodities have been used to achieve several objectives and to reach different populations. For example, the commodity donation programs that began in the 1930's were the antecedents of the current Food Stamp 1-1 Program, now the nation's largest food assistance program funded at over $17 billion in FY 1991 and serving over 22.6 million people. Schools first began receiving food donations in the 1930's. However, it was not until the adoption of the National School Lunch Act of 1946 that the role of food assistance in the schools was formalized and provided on a regular basis. Currently, the NSLP is by far the largest commodity donation program. However, there are another 12 programs that serve such widely divergent populations as charitable institutions, summer camps, and Indian reservations. Debate over whether to provide food assistance in the form of increased purchasing power (e.g., cash, food stamps, or letters of credit) or in the form of commodities has been a part of the policymaking process for these programs since their earliest days. The initial move tov/ard providing eligible food aid recipients with redeemable coupons instead of donated foods began in May, 1939 when the U.S. Department of Agriculture's first experimental food stamp program was initiated in Rochester, New York. & Still, it was not until many years later, in August 1964 that a full-scale, nationwide Food Stamp Program was established. In the case of school food service programs, a combination of cash assistance and commodity assistance has been provided since the beginning of the NSLP. During the early years of the program, commodity donations played a more important role than they do now. In FY 1965, commodities accounted for over two-thirds (68 percent) of Federal support.^ Over time, the relative importance of commodity assistance to schools has declined, although it continues to be important to industry in surplus removal and market stabilization. Still, donated commodities remain a significant source of support to school food service programs throughout the nation. In SY 1990/91 schools participating in the NSLP received donated foods valued at $624 million or about 17 percent of total Federal assistance to the program. Donated commodities are classified as "entitlement" or "bonus". Entitlement commodities are distributed at a legislatively determined rate based on the number of lunches served. In SY 1990/91, school districts were entitled to receive commodities valued at 14 cents for every lunch served that met USDA requirements. Since SY 1981/82, the entitlement rate has increased at an average annual rate of 2.8 percent. Bonus commodities are made available as condition of crop surpluses warrant. Bonus commodities were first distributed in FY 1978. With large surpluses of dairy products in the mid-1980's, bonus commodities grew rapidly reaching $440 million in FY 1987 and accounting for 49 percent of the value of all donated commodities through the NSLP that year. More recently bonus commodities have been distributed in far smaller amounts. In SY 1990/91, $83 million of bonus commodities were distributed. U U.S. Congress, Senate Committee on Agriculture, Nutrition, and Forestry, The Food Stamp Program: History. Description. Issues, and Options. April 1985, p.4. ^ Congressional Budget Office, Feeding Children: Federal Child Nutrition Policies in the 1980's May 1980, p.50. 1-2 Trends in the level of federal assistance to school feeding programs over the period 1969-90 are shown in Table 1-1. As the nation's school systems consolidate and grow larger, as the technology and resource requirements of meal preparation evolve, and as the nation's eating habits change, there are new and different requirements for the types, forms, and unit sizes of foods that schools can most effectively use as well as changes in the nature of the distribution system through which these foods are made available to the schools. Within the farm sector, however, the need for price stabilizing actions in commodity markets has remained, though with the exception of the early 1980's, the farm economy has been far healthier than it was during the Great Depression. For mo<st of the basic field crops, such as wheat and feed grains, and for milk, price and income stabilization has been sought through use of commodity programs tailored to the individual commodities. In the case of field crops, this has been achieved through use of nonrecourse loans, deficiency payments, and acreage diversion measures. Milk prices have been supported through purchases of manufactured dairy products by the Commodity Credit Corporation of the USDA. For the so-called specialty crops, however, including fruits and vegetables, there have been no comparable programs, aside from the more limited authorities provided under the Federal Marketing Order Program. As a result, producers of these crops have become somewhat more reliant on USDA purchases under the Commodity Donation Program as a means of market stabilization than have producers of basic commodities. Over the years, there have been changes in the food donation program in response to the changing needs of the participating schools. Yet, in the view of some school food administrators, these program changes have not kept pace with their changing needs. As a result, the food donation program has come under increasing criticism. As far back as 1974, the USDA began evaluating alternatives to the commodity donation program including use of cashouts, vouchers, and price discounts.-*/ This was followed in 1978 by a Congressional directive to conduct a pilot study of an all-cash system. The inconclusive results of these studies eventually led to another Congressional directive. As part of the FY 1981 Agricultural Appropriations Act, the Department of Agriculture was directed to conduct an extensive 3-year study of two alternatives to the commodity donation program - - an all cash approach and CLOC. B. Program Operations As a prelude to discussion of the principal focus of this evaluation, the major operating elements of the existing Commodity Donation Program and the CLOC program are described in this section. V Office of Planning and Evaluation, USDA. "Implications of Discontinuing USDA Commodity Acquisition Domestic Activities," January 9,1974 (unpublished), and Office of Planning and Evaluation, USDA, "A Study of Alternatives to Commodity Distribution Programs," September 25, 1974 (unpublished). 1-3 Table 1-1 Federal Support for School Food Proerams FY 1969-FY 1990 Commodities Fiscal leax Total Cash Assistance 1/ Commodities Assistance Total Federal Support As percent of Total Entidement Bonus Total Federal Support percent 1969 209.2 272.0 272.0 481.2 56.5 1970 311.0 265.2 265.0 576.0 46.0 1971 551.6 277.3 277.3 828.9 33.5 1972 763.6 312.1 312.1 1,075.7 29.0 1973 916.7 331.0 331.0 1,247.7 2C.5 1974 1,144.5 316.1 316.1 1,460.6 21.6 1975 1,375.1 423.5 423.5 1,798.6 23.5 1976 1,605.2 418.6 418.6 2,023.8 20.7 1977 1,718.9 540.8 540.8 2,259.7 23.9 1978 1,989.5 485.3 57.<i 542.9 2,532.4 21.4 1979 2,214.7 675.3 69.e» 744.9 2,959.6 25.2 1980 2,567.2 765.5 139.C1 904.5 3,471.7 26.1 1981 2,712.3 578.9 316.3\ 895.2 3,607.5 24.8 1982 2,502.7 426.2 330.?t 757.0 3,259.7 23.2 1983 2,745.6 426.8 374.1 800.9 3,546.5 22.6 1984 2,871.7 440.5 386.51 827A 3,699.1 22.4 1985 2,957.7 456.0 345.2 801.3 3,759.0 21.3 1986 3,120.9 445.7 376.2 821.9 3,942.8 20.8 1987 3,244.0 448.5 439.C 888.2 4,132.2 21.5 1989 3,398.5 466.3 347.4 813.7 4,212.2 19.3 1990 3,519.2 471.7 291.5 763.1 4,282.3 17.8 y Includes cash assistance through the National School Lunch Program and the School Breakfast Program; excludes assistance through the Special Milk Program. Source: Food and Nutrition Service, U.S. Department of Agriculture. V Commodity Donation Program Under the existing program, commodities purchased by the USDA are made available to eligible participating agencies, in this case school systems participating in the NSLP. The value of the commodities these school systems receive is determined by two factors. First, a fixed per meal value is legislatively determined each year as a basis for determining the value of "entitlement" commodities schools are eligible to receive. In SY 1990/91, this was set at 14c per reimbursable meal, with the number of reimbursable meals based on the preceding year, with an adjustment for differences between actual and predicted values for the preceding year. In total, entitlement commodities valued at $545.1 million were distributed in SY 1990/91. Second, additional foods, called "bonus" commodities, are made available to schools depending on market conditions and the need to purchase commodities that are in a state of temporary surplus. Given the variability in the size of Federally-owned surplus stocks, the value of bonus commodities has varied greatly from year-to-year and in SY 1990/91 was substantially lower than it has been for several years. Preliminary estimates indicate that only about $83 million in bonus commodities were distributed to schools in SY 1990/91. This is less than 20 percent of the value of bonus commodities donated in FY 1987. When large quantities are available, bonus commodities are made available to school systems in whatever amounts they can use without waste. When these commodities are available in more limited quantities, as they were in SY 1990/91, they are allocated among the participating school districts. A relatively wide range of products is made available through the existing program. A list of foods distributed through the program in SY 1990/91 appears in Table 1-2. Although the composition varies somewhat from one year to the next, many of the same foods are offered nearly every year, though in varying quantity depending on market conditions. In SY 1990/91, over 80 different foods processed from 32 commodities were donated to schools. Several of the foods are offered in different forms (e.g., canned and frozen green beans). Donated foods are provided to the schools in institutional-size units, for example, canned goods are normally packaged in cases of six-# 10 cans, while ground beef is distributed in frozen 6-pound or 9-pound chubs in boxes weighing 36 pounds. Most foods donated through the program have been processed to the minimum extent necessary to make them usable for this purpose. As a result, many are more appropriately characterized as ingredients than as final food products. Nonetheless, in response to recipient agency calls for greater convenience in preparation, there has been some shift toward the purchase of more highly processed foods, particularly those containing poultry and red meat. For example, in addition to donating blocks of frozen ground beef, the USDA also donates ground beef in pre-formed patties and ground beef combined with vegetable protein. In addition, in SY 1990/91 the Department experimented with the purchase of alternative formulations of extra lean ground beef in pre-formed patties. Most fruit and vegetable products are purchased once and sometimes twice during the year, generally coinciding with the pack season. Staple foods such as flour, rice, and vegetable oil are purchased monthly throughout the year. Meat and poultry products are normally purchased weekly beginning in the late summer and continuing through January or February. This allows USDA to adjust purchases in response to market conditions. Nearly all donated foods are purchased commercially through competitive bid and are 1-5 Table 1-2 Food Donated to School Districts Through The Commodity Distribution Program. School Year 1990/91 Fruit and Vegetables Green beans, canned Mixed vegetables, frozen Green beans, frozen Apples, fresh Vegetarian beans Apple slices, canned Refried beans Applesauce Sweet corn, canned Apricots, canned Sweet corn, frozen Cherries, frozen Green peas, canned Peaches, canned Green peas, frozen Pears, canned Potato rounds, frozen Pears, fresh French fries, frozen Pineapple, canned Sweet potatoes, canned Mixed fruit Tomatoes, canned Blackberries, frozen Tomato paste Raspberries, frozen Raisins Date pieces Meat, Poultry. Seafood, and Dairy Products Cut-up chicken, frozen Beef patties, extra lean Diced chicken, frozen Luncheon meat, canned Bulk chicken, chilled Ground pork, frozen Whole turkeys, frozen Cooked boneless ham, frozen Bulk turkey, chilled Canned tuna Turkey roasts, frozen Process cheese Ground turkey, frozen Cheddar cheese Whole eggs, frozen Mozzarella cheese Ground beef, frozen Nonfat dry milk Beef patties, frozen Butter Beef patties w/VPP, frozen Grain. Oilseeds. Nuts and Legumes Flour Milled rice Grits Brown rice Macaroni Rolled wheat Rolled oats Vegetable oil Peanut butter Soybean oil Peanut granules Salad dressing oil Roasted peanuts Vegetable shortening Spaghetti Dry beans Almond butter Cornmeal Honey Source: Food and Nutrition Service, USDA, "Order Summary Report for 7/1/90 through 6/30/91," (unpublished). 1-6 purchased expressly for the purpose of donation to participating school systems. Contrary to common belief, most donated foods do not originate from Federal stocks. The most important exception to this is the large volume of manufactured dairy products that has been a source of donations in recent years, though some dairy products have also been purchased commercially for purposes of donation. The USDA arranges for the delivery of donated foods to central delivery points within each State. Responsibility for the further storage and distribution of these foods, once they arrive at these central delivery points, has been delegated by the Secretary of Agriculture to State Distributing Agencies (SDA's) designated by the governor of each State typically the State Department of Education or Agriculture. Because there are more than a dozen different food donation programs serving widely divergent purposes, some governors have divided program responsibility among more than one SDA within their States. Within certain broad parameters, SDA's have been free to conduct their responsibilities to best suit the needs of their respective States. As a result, a 1988 study of State commodity distribution systems found vast differences in methods of operation and levels of performance.^ Some States contracted with commercial distributors to store and distribute donated commodities. Others operated their own warehouses and delivery fleets. Still others arranged for central storage but left it to the individual school systems to arrange for pick-up or delivery. Some States negotiated a wide array of processing agreements through which donated foods were further processed before delivery to the schools. Other SDA's had few or no processing agreements. Some States levied an assessment against participating schools to cover at least a portion of the cost of their distribution program while other states absorbed the full cost. The Commodity Distribution Reform Act of 1987 sought to improve the efficiency and effectiveness of this distribution system by requiring each SDA to evaluate its current system and determine whether a commercial alternative would be superior. Where the commercial alternative is found to perform better, States are required to convert to a commercial system. Regulations implementing this law are now in effect. While there have been some changes in SDA operations since the adoption of this legislation, substantial differences remain among the systems. Thus, the existing commodity donation program is uniform in some respects but not in others. Administrators of the Commodity Donation Program and those programs in other agencies that provide supportive services have made numerous improvements in the program over the past 5 to 10 years. Improved products (including reduced fat and salt content), improved labeling and product identification, the adoption of unitized deliveries, improved packaging, increased product variety, greater attention to nutrient analysis, and improved scheduling of deliveries are examples of these changes. Thus, a number of the criticisms described below that gave rise to the development of CLOC have been addressed in the intervening years. V Abel, Daft & Earley, A Study of the State Commodity Distribution Systems, prepared for FNS, USDA March 1988. 1-7 The Commodity Letter of Credit Origin of the Concept. The CLOC approach was designed in response to particular criticisms of the commodity donation program. More specifically, the donation program was faulted for the following reasons: • Commodities were frequently donated in forms that could not be conveniently used by school feeding programs, conventionally referred to as "school food authorities" (SFA's). As a result, meal preparation from these foods often required substantial labor inputs. Whole turkeys, 36 pound cartons of frozen ground beef, and 50 and 100 pound bags of flour are examples of the forms that were considered inconvenient. • Schools lacked control over the time of delivery, making inventory control difficult and adding to the cost of storage. Lack of control also disrupts menu planning which is important to the efficient scheduling of procurement. • Because donated commodities satisfy only about 20 to 30 percent of their food needs, SFA's are required to maintain a dual system, one for commercial purchases and another for commodity donations. Purchased foods and donated foods are different in many respects: they are commonly distributed through different systems: as they arrive at the schools the foods are in different forms; their deliveries cannot be easily coordinated; and there are often variations between the two in quality. • Some of the donated commodities could not be effectively used due to their unacceptability by school food participants. It should be noted that these low-preference foods represent a very small share of the total value of foods distributed, probably accounting for no more than 2 or 3 percent. When the CLOC program was initially developed, therefore, it was designed largely for the purpose of correcting these shortcomings while continuing to pursue the principal objectives of the present program. The description of the CLOC program that follows is based on the program as it was initially designed and implemented. Modifications to the program which were implemented in SY 1990/91, and which are the subject of this evaluation, will be discussed later in this Chapter. CLOC Issuance and Use. Under the CLOC approach, school systems receive cash allocations equivalent to the value of the donated commodities that they would have received through the donation program. This allocation is equivalent to the entitlement values described earlier. These cash allocations are made to CLOC schools on a quarterly basis, beginning in July, for use in buying foods authorized under the CLOC program. Following the pattern of USDA purchases, CLOC authorizations are proportionately larger in the firs* half of the school year. Therefore, the first two allocations each account for about 36 percent of the total with 21 percent and 7 percent distributed in the third and fourth quarters, respectively. 1-8 The treatment of bonus commodities in the CLOC program has been a controversial issue, resulting in their being treated in different ways in different years. During the early and mid-1980's when the Commodity Credit Corporation held large stocks of butter, cheese, and nonfat dry milk that they were distributing as "bonus" commodities, the USDA argued that CLOC systems should be required to accept these foods in their donated form. Thus, with the exception of SY 1984/85 and part of SY 1982/83, CLOC schools have received bonus commodities in donated form. When the USDA purchases a particular commodity for donation, a commodity letter of credit is issued to participating schools authorizing them to use a portion of their CLOC funds to buy products containing the commodity USDA has purchased. The share of funds they are authorized to use is proportional to the share of total entitlement funds for all schools that USDA has used in making the purchase. Schools may use their CLOC funds to buy a wide range of foods containing the designated commodity. As initially implemented, participating schools were granted full credit against their CLOC funds if the food they purchased was 90 percent or more, by weight, of the commodity purchased by USDA or was a product that USDA had once purchased. Additionally, schools were granted "partial credit" against their CLOC funds if the donated commodity was used as an ingredient in a processed product. The amount of the partial credit was based on the share of total product weight accounted for by the ingredient. As the program was initially designed, CLOC's were issued when the USDA announced that they had made a purchase. For those foods that are purchased on a continuing basis by USDA at least for a portion of the year (such as flour or ground beef), CLOC's are aggregated and issued on a quarterly basis. To provide CLOC schools with greater control over the timing of purchases and deliveries, as originally implemented these schools received advance notice through CLOC alerts at the time USDA announced its intention to make a purchase. Then, when USDA announced its purchase, CLOC's were issued. The CLOC's required that both purchases and deliveries be completed by specified dates. The final purchase date was generally set to coincide with the end of the shipping period for delivery to State Distributing Agencies. The final delivery date was either the end of the school year or the end of the second quarter of the following year, depending on when the CLOC was issued. To provide CLOC schools with at least as much flexibility as offered by the existing program, these schools were given an opportunity to amend their CLOC issuances in the following ways. (1) They were permitted to refuse CLOC's that they did not want and receive substitute CLOC's in their place. Substitute CLOC's were offered from commodities that had been offered as "bonus" commodities (excluding dairy products). They were allowed to trade CLOC's with others CLOC school districts as long as the dollar amounts of the trades were equal. And (3) they were permitted at the beginning of the fourth quarter of each year to consolidate unused portions of CLOC's within commodity categories (e.g., poultry), as long as the unused portions did not exceed 10 percent of the original dollar value, and to receive a "rollover" CLOC that could be used to purchase any of the commodities for which CLOC's had been issued that year. 1-9 Since its inception in 1982, the CLOC program has been administered by an independent contractor, Virginia Polytechnic Institute and State University. In consultation with USDA's Food and Nutrition Service, (FNS), this contractor determines the levels of entitlement for participating school systems, calculates and issues CLOC's, receives and approves documentation in support of CLOC purchases, and generally oversees administration of the CLOC program. C. Results of Past Evaluation The CLOC program was originally designed and implemented in response to a Congressional directive in the Agriculture Appropriation Act for Fiscal year 1981 (PL 96- 528). The Act directed the USDA to conduct a demonstration study comparing results of the existing commodity donation program to the results of two new alternatives-the commodity letter of credit and an all cash approach. The study was conducted by Abt Associates. ^ The demonstration began in SY 1982/83, though information was collected for SY 1981/82 to be used as a baseline in the evaluation. As a result of a change in the treatment of bonus commodities midway through SY 1982/83, the information collected that year was of little, if any, value for purposes of the evaluation. As a result, the first evaluation year was SY 1983/84. During this year, all schools participating in the demonstration received cash or CLOC's in lieu of entitlement commodities and continued to receive bonus commodities. For the following year, SY 1984/85, Congress extended the demonstration and required that the CLOC and cash systems receive both entitlement and bonus benefits in the form of CLOC's and cash, respectively. This was required so that the alternative approaches could be compared in their "pure" forms. A sample of 96 school systems from 29 States was selected for the study. The sample was purposefully drawn in the sense that it was required that all seven FNS regions were to be represented, each by at least three States, and that these States were to represent a diversity of agricultural, transportation, and food distribution conditions. For purposes of selecting individual systems to participate in the demonstration, all SFA's within each of the selected states were grouped according to size, poverty level, and participation in feeding programs. Then, within each state, six SFA's were randomly selected from that group that most closely represented the predominate characteristics of the state. From this number, three SFA's were randomly selected from within each state and were then randomly assigned to one of the three options under study ~ cash, CLOC, and the existing commodity donation program. Very large school systems were selected somewhat differently since there was a desire to avoid affecting individual states too severely. The composition of the final sample was: 32 current program SFA's, 31 CLOC SFA's, and 33 cash SFA's. An extensive volume of data was collected from the participating school systems over the period of study. In addition to collecting information on a wide variety of operational • For an overall description of the study and its findings see: St. Pierre, Robert G., et. al., Evaluation of Alternatives to Commodity Donation in the National School Lunch Program: Final Report. Abt Associates, Volume 6, 1986. MO characteristics, detailed records were collected on all food acquisitions by item, by month and for end-of-year inventories throughout the period. An assortment of performance measures for school systems in each of the three program options were compared for the two study years (SY 1983/84 and SY1984/85) as were changes between the study years and the base year (SY 1981/82). Findings from this evaluation fill several volumes. In brief, it was concluded that: • During the demonstration, a per-lunch entitlement of 12 cents per NSLP meal was uniformly applied to school districts in all three treatments. However, the entitlement does not represent all of the costs associated with the purchasing and distribution of donated commodities (e.g. administration, storage, and transportation). As a result, the market value of the commodities provided to the control school districts was about 3 cents more per meal than the Federal subsidy provided to the cash and CLOC districts. CLOC Districts increased their own outlays for food by about 1 cent per meal to compensate for the 3 cent reduction in the value of their commodity subsidy. Thus, if the market value of the entitlement had been the same for all three groups, there would have been no difference among the treatments in the market value of food acquired. • While the per lunch value of CLOC benefits was set at a uniform level for all schools participating in that treatment, schools participating under the existing program were found to have received widely different levels of benefits (adjusted to a per lunch basis). In 1984/85, the value of entitlement was found to range from 9e/lunch to 17e/lunch. Since schools in the existing program had the right to refuse any commodities that they did not want or to accept more commodities if available, at least part of this variation is thought to have resulted from their exercise of this right. And, those schools that accepted these commodities could have received more than their entitlement. However, there was no information with which to test this hypothesis. • Schools receiving CLOC's reduced their inventory levels sharply, 50% during the first year of participation. Though donations accounted for only 30 percent of food acquisitions in the base year, they accounted for over 80 percent of the foods in inventory. • By purchasing their foods in a more highly processed form, schools receiving CLOC spent about 6 percent more of their food dollars on processing and distribution and a correspondingly lesser amount on commodities. • CLOC schools used less of some commodities (beef, chicken, pears, and sweet potatoes) and more of others (field corn and cottonseed). • CLOC schools did not use their CLOC certificates to buy locally produced fresh products to any extent. Ml • CLOC appeared to have no significant effect on labor costs, though this could have been due to the relatively short-term nature of the study. Storage and distribution costs were reduced by about 50 percent. • There was no di'.-ernable effect on student participation in the program. • The School Food Authorities who operated under CLOC were strongly supportive of the program and the increased flexibility it provided. To policymakers in the USDA, results of the 1981-85 demonstration seemed inconclusive. While the CLOC program yielded a few small benefits, it also resulted in slightly smaller volumes of some commodities being removed from the market, thereby reducing the program's effectiveness as a means of price support. Absent more conclusive evidence of the benefits of CLOC, the Administration concluded that the program did not merit continuation and recommended that authority for the program be allowed to expire, This position was supported by several commodity groups and by the American School Food Service Association. Most of the SFA's that had participated in the CLOC demonstration came to a different conclusion, however. They were particularly pleased with the increased flexibility in procurement and scheduling that the CLOC program had provided. And while they might have sacrificed some dollar support, the increased autonomy more than compensated them for this loss. The SFA's that had received cash in lieu of commodities were similarly pleased with the effects of their option. As a result of their strong support for these options, 59 of the 63 SFA's that had received CLOC's or cash as part of the demonstration joined together in forming a new organization, Schools for an Alternative to Commodities. In the years since the completion of the original demonstration, this organization has successfully lobbied for these SFA's to be allowed to continue receiving CLOC's and cash. This organization has also sought, thus far unsuccessfully, for the CLOC program to be extended nationwide. Other interests, namely commodity organizations and operators of the present program, have remained opposed to the CLOC approach. While program operators and school food service managers are concerned with administrative and monitoring issues, the opposition from agricultural interest groups is based on several concerns relating to the effectiveness with which CLOC removes targeted commodities from the market and, thereby, provides price support. These concerns grow out of certain operational procedures that were adopted under the program's original design. The most important of these are as follows: (1) Level of Specificity. CLOC's are issued by commodity name, e.g. "apples," "tomatoes", or "peaches." There is a concern that in issuing CLOC's for generic commodities (e.g. "peaches" rather than "cling peaches"), the Department loses its ability to target purchases on more narrowly defined commodities. (2) Crediting Procedures. As noted above, SFA's were granted substantial discretion in how they used their CLOC's. They could apply the full cost 1-12 of the food against their CLOC funds if it was primarily composed (90 percent by weight) of the commodity indicated. And, if the commodity accounted for less than this, partial credit was allowed based on the share of total weight accounted for by the commodity. Commodity interests argued that this procedure resulted in a substitution of processing value for commodity value. As a result, a reduced volume of commodity was removed from the market per dollar of Federal program expenditure. (3) Timing of Purchases and Deliveries. Critics of the CLOC program also argued that the latitude given these SFA's in the timing of purchases and deliveries results in a more dispersed pattern of purchases. In comparison with the large, concentrated purchases of USDA, they argued that this is a less effective means of registering the additional demand that is supposed to be associated with the purchase and, therefore, it is less effective in supporting farm prices. (4) CLOC Adjustments. In the interest of giving CLOC SFA's an opportunity to make the maximum possible use of their CLOC's and in order to mirror the commodity program as closely as possible, they had been permitted to make certain prescribed adjustments in CLOC's issuances, as noted earlier. This included (a) trades of equal value among CLOC SFA's, (b) the reissuance of CLOC's for substitute commodities, and (c) the consolidation of unused CLOC funds in "rollover" CLOC's. Critics of the CLOC program argued that the use of substitute and rollover CLOC's reduced the donation program's effectiveness by altering the composition of commodities removed from the market and by Allowing purchases to be made after the period when additional demand in the market was most needed. (5) Domestic Origin. When USDA buys foods for donation, it requires that the foods and the commodities used in manufacturing the foods be entirely of domestic origin. Though the same domestic origin requirements apply to the use of CLOC's, critics of the program charged that individual school districts are less able to monitor and enforce the requirement than is the U.S. Department of Agriculture. And, to the extent CLOC funds are used to purchase imported foods, it represents a further erosion of support for domestically produced commodities. D. Modifications of the Commodity Letter of Credit The Conference Report accompanying the Child Nutrition and WIC Amendments Act of 1989 provided for a test of modifications to CLOC that might address the concerns of commodity interests without negating the program's advantages to SFA's and without causing the program to become prohibitively expensive to administer. On the basis of this authority, and as a result of his personal interest in the topic, former Secretary of 1-13 Agriculture Clayton Yeutter directed FNS to design modifications in the CLOC program and to evaluate their effects. To identify appropriate changes, FNS officials met with representatives of agricultural interests in January 1990 and with school food directors from CLOC SFA's in March 1990. Out of these meetings and subsequent deliberations within FNS, the following modifications in the CLOC program were developed for implementation in SY 1990/91. • Commodity Specificity. The degree of commodity specificity for which CLOC's are issued is to be determined by USDA on a case-by-case basis. If the Department determines that a more narrowly defined commodity market is to be targeted, it will advise CLOC program administrators to specify the CLOC accordingly. However, unless a narrower market is being targeted, CLOC's will continue to be issued for generic commodities. It should be noted that only two commodities were affected by this change in SY 1990/91 - pears and beef. In the case of pears, separate CLOC's were issued for "Bartlett pears" and for "fresh winter pears." And, it was determined that beef CLOC's were to be exclusively used for the purchase of beef from culled dairy cows. With these exceptions, CLOC's issued in SY 1990/91 were identical to those of earlier years. • Full/Partial Credit. The crediting of CLOC purchases is now based on the farm equivalent weight V of the targeted commodity per dollar spent. Full credit is given when the farm-level equivalent weight per dollar spent equals or exceeds that of USDA purchases. Fresh food purchases also receive full credit. Partial credit is given for foods not qualifying for full credit and is based on the ratio of farm-equivalent weight per dollar spent for the purchased food to the farm-equivalent weight per dollar spent for the USDA donated food. CLOC program administrators have provided each of the participating SFA's with a list of foods qualifying for both full credit and partial credit, including in the case of the latter the percentage credit allowed. As additional qualifying foods are added, these lists are updated. This modification has substantially changed the proportion of the cost credited to CLOC for some processed foods. For example, under the old procedure apple juice in individual servings qualified for full credit since it is made up entirely of apple juice. Under the new procedure, only 70 percent of the cost of apple juice in this size container would be credited to an apple CLOC. • Timing of Purchases and Deliveries. Under the modification, the time periods within which CLOC purchases are to be made and deliveries received have generally been accelerated and shortened to better match the schedule of USDA purchases and deliveries. CLOC's are now issued at the time USDA advertises for bids. They had formerly been issued at the time V Farm-equivalent weight is the weight of a raw agricultural commodity at the farm gate that is required to produce a given unit of a product containing that commodity. 1-14 contracts were awarded. Purchases must now be completed by the time contracts are awarded, though a minimum of 30 days to make commitments to purchase is granted. Deliveries are now required by the final shipping date of the USDA purchased foods. For those foods that USDA purchases on a continuing basis, quarterly CLOC's continue to be issued. • Trades. Rollovers, and Substitutions. CLOC trades continue to be permitted, though unlike in the past, they may now be of unequal value. The consolidation of remaining balances in rollover CLOC's and the replacement of refused CLOC's with substitutions are no longer permitted. , • Domestic Origin. For purposes of tightening the domestic origin requirement, the modification requires all distributors and manufacturers supplying CLOC SFA's to certify that the products they supply were grown and produced in the United States. Also, in submitting their CLOC reporting forms, participating SFA's are required to certify that all foods claimed for CLOC credit are, to the best of their knowledge, of domestic origin. These modifications were implemented at the beginning of SY 1990/91. Their effects on CLOC program performance, particularly those elements of performance that were of concern to critics of the program, are the principal subject of this evaluation. 1-15 IL OBJECTIVES OF THE EVALUATION A. Issues to be Examined There are two central issues to be examined in this evaluation, both of them related to the modification of the CLOC program adopted in SY 1990/91. They are: 1) The effect of these modifications on the performance of the CLOC program relative to the performance of the existing Commodity Donation program. 2) The effect of these modifications on the performance of the CLOC program as it is currently operated relative to its performance prior to the modifications. All of the adopted modifications have been made in response to criticisms of the CLOC program for deviating in various ways from the existing Commodity Donation program. Each change has been designed to bring the operation of the programs, if not their performance, into closer conformity. Therefore, from the outset of the study the two systems were expected to have at least moved in the direction of fewer differences than have been observed in the past. The outcome with respect to changes in the performance of CLOC SFA's compared to their combined performance prior to the adoption of the modifications was harder to anticipate. Because there has been very little, if any, criticism of CLOC from within its own ranks, none of the adopted changes stemmed from internal criticism of the program. The Congressional authorization for modifying the CLOC program recommended that changes in the program not result in "the pilot sites being overburdened with new complex procedures or rules or modifications that negate the major benefits that the school food service directors operating Cash and CLOC programs believe are offered by these alternative programs."^ Given that some modifications have resulted in relatively significant procedural changes, careful attention was given to the effect of these changes. Most performance measures are dictated, directly or indirectly, by the nature of the principal criticisms that led to the modifications that were adopted. While a number of different performance measures were highlighted, two are of paramount concern from the viewpoint of CLOC critics. They are: (1) the farm-equivalent weight of individual commodities removed from the market and (2) the timing of these market removals. There may be other operational definitions of agricultural impact or factors influencing this impact that are not captured by these two quantitative measures. From the standpoint of participating CLOC SFA's, the key measure of performance is one that is difficult to quantify: operational flexibility. For CLOC SFA's, maintaining sufficient flexibility in choice of product and timing of purchases and delivery is at the forefront of their concerns. It should be noted that this study will not address issues related to administration of the program at the State or National levels. If CLOC were to be expanded beyond its current pilot basis, there are several administrative issues that would require attention. They include: the identification of the roles to be played by FNS headquarters, FNS regional offices, the SDA's, AMS, and ASCS; determining the labor requirement of each of these * Congressional Record. H6862, October 10, 1989. II-1 IU units; developing procedures to administer CLOC on a larger scale (e.g. issue checks, issue CLOC's, monitor and supervise CLOC use); and evaluate the effect of CLOC on the other commodity donation programs. B. Research Questions To move the evaluation from these general issues to analysis, two additional steps were taken. First, these general issues were translated into a list of 14 study objectives. They are listed in the left-hand column of Table II-1. Each objective states a particular comparison or determination that is to be made in the course of the evaluation. Some of these objectives are interdependent in that they are designed to help explain the outcomes of other objectives. While there is no rigid hierarchy in the ordering of these objectives, they are generally listed from most to least important in terms of evaluation priority. Having identified each of the principal objectives, each objective was translated into one or more research questions. These questions are listed in the right-hand column of Table II-1 alongside their associated objective. Most of these research questions are precisely framed and are designed to be empirically answered. Some of the questions toward the end of the list are more subjective and will result in less precise answers. Findings for each research question are reported in Chapter V. Table H-l Objectives Research Questions i. Determine the extent to which CLOC modifications result in a difference in the quantity of commodity removed from 1. Do CLOC SFA's remove the same weight of each commodity the market compared to the *xwtint food donation per LEQ as Commodity SFA's? program. II. To the extent there are differences in the quantity of commodity removed, determine the extent to which it is 2. Is there a difference, on the basis of product value, in the mix due to the purchase of more highly processed products. of products removed from the market within each conurodity between CLOC SFA's and Commodity SFA's? 3. Is there a difference in the proportion of the value of the food products derived from each commodity that is attributable to its raw commodity value for CLOC SFA's as compared to Commodity SFA's? in. Where there are difference-, in &: quantity of commodity removed, determine the rxtent to which it is due to CLOC 4. What proportion of the value of CLOC's issued are not spent? trades or failure to sper d the face value by CLOC SFA's or to Commodity SFA's not receiving delivery of the 3. What proportion of the value of their entitlement do commodity. Commodity SFA's receive? IV. Determine thi.1 extent to which there are differences in the value of benr.its under CLOC and Commodity Donation 6. Is there a difference in the value of CLOC issuance per LEO and ii.r extent to which any differences are due to and the value of commodity donations per LEQ? program design versus program execution. 7. How does the weight of commodity acquired per dollar of v. Determine the extent to which modifications in CLOC CLOC issuance in SY 1989-90 (before modification) compare have altered the way in which they are used by CLOC to the weight per dollar of CLOC issuance in SY 1990-91 SFA's. (after modification)? II-2 n Table II-1 (continued) VI. Compare the timing of purchases and deliveries CLOC and Commodity programs VII. Determine the extent to which CLOC has affected Urn use of labor. VIII. Compare the effect on rates of participation in school feeding programs of the modified CLOC program and the Commodity program. DC Determine CLOC SFA managers' views of the effects of the CLOC modifications. X Determine the effect of CLOC on other food donation programs. XI. Determine USDA commodity procurement procedures and implications for the CLOC program. XII. Determine why USDA purchase prices are often below commercial prices. XIII Determine the feasibility of policing the requirement that CLOC's only be used to purchase foods that are grown and processed in the U.S. XIV. Determine labor requirements to administer the modified CLOC program. 8. How does the proportion of the value of CLOC issuance not spent in SY 1989-90 (before modification) compare to the proportion of the value of CLOC issuance not spent in SY 1990-91 (after modification)? 9. Hew does the proportion of the value of CLOC issuance that was refused, rolled over, ti aded or not spent in SY 1989-90 compare to the proportion that was traded or not spent in SY 1990-91? 12. How does the pattern of use of CLOC's by month following date of issuance in SY 1989-90 compare to the pattern of use in SY 1990-91? 10. How does the timing of CLOC purchases compare to the timing of USDA purchases? 11. How does the timing of CLOC deliveries to CLOC SFA's compare to toe timing of donated commodity deliveries to Commodity SFA's? 13. How do changes in labor cost, labor hours, and food costs per LEQ between SY 1989-90 and SY 1990-91 for CLOC SFA's compare to changes in the same measures over this period for Commodity SFA's? 14. How do the rates of labor use by CLOC SFA's during the earlier demonstration (SY's 1983-84 and 1984-85) compare to the rates of labor use for these same SFA's in 1990-91? 15. How does the change in rates of participation in school feeding programs between SY 1989-90 and SY 1990-91 in CLOC SFA's compare to the change in these measures for Commodity SFA's? 16. How do CLOC SFA managers view the effect of CLOC modifications on Hie popularity and nutritional content of meals served and on the oost of program administration? 17. What effect would (he adoption of the CLOC program nationwide have on the operation of other commodity donation programs? 18. What criteria and procedures are followed by the USDA in determining which commodities to purchase and the volume and timing of the purchases? 19. Why are the prices paid by USDA for the commodities that they buy frequently lower than commercial prices available to SFA's? 20. To what extent are CLOC's used to purchase foods that are not of domestic origin? 21. How has the modification of the CLOC program affected Virginia Tech's cost of program administration? 22 What have been the labor requirements of USDA staff in administering the CLOC program during SY 1990-91? II-3 IE m. EVALUATION METHODOLOGY A. Evaluation Sample Information was collected from 50 SFA's in 23 States throughout the nation for this evaluation. Half of these SFA's participated in the CLOC program and half participated in the Commodity Donation program. While the evaluation sample includes districts from various parts of the country with varying enrollment sizes in both rural and urban settings, the sample was not designed to be nationally representative. Data necessary for an extensive comparison of the study sample and national totals were not available. The sample was formed in two stages. When the CLOC program was first established at the direction of the Congress in 1981, a nationally representative sample of SFA's participating in the NSLP was drawn to participate in the demonstration that followed. Of the 32 SFA's in that sample that participated in the CLOC program, 25 were still in the program in SY 1990/91 and, as a result, were in the sample for this evaluation. Having participated in CLOC since SY 1981/82, these SFA's are familiar with the program and therefore better prepared to make adjustments required by the modifications than would SFA's that had not participated in the program. The sample for the 1981 study was drawn in a 4-step process that began with the purposeful selection of 29 States representing geographic, agricultural, and distribution system diversity. Within each of these States, SFA's with enrollments greater than 25,000 were isolated and the remaining SFA's were categorized by participation in lunch or lunch and breakfast programs, size of enrollment, and poverty level. From this pool, six SFA's were selected from that category that was most typical of the State with two assigned to each of three treatments (CLOC, Cash and Commodity). The very large systems were categorized in a similar manner using cross-State pools and six additional SFA's were randomly selected from each of these pools. At the direction of Congress, some volunteer SFA's were also included in the pools that had been formed by this selection process. Finally, three SFA's were randomly selected from each of the within-State and cross-State pools and were randomly assigned to one of the three treatment groups. The sample that was drawn in 1981 over-represented large SFA's and under-represented small SFA's. Of the 96 SFA's in the original sample, 7 percent had enrollments of 25,000 or more compared with 1 percent nationally. This was done because small SFA's account for a very small share of total enrollment. Thus, in terms of numbers of students and overall school food service activity, large SFA's have a proportionately larger impact than small SFA's. This over-representation of larger SFA's is continued in the sample used for this evaluation, though it is greatest for those SFA's with enrollments of 5,000 to 24,999 rather than systems of 25,000 or more. As in the earlier study, the sample used for this evaluation slightly over-represents the share of lunches that are full-priced (61.8 percent in the sample versus 53.7 percent nationally) and under-represents the share that are free or reduced price. A larger difference between the sample and the national universe is in the share of SFA's with a School Breakfast Program (SBP). Of the SFA's in the sample for this evaluation, 56.0 UI-1 /? percent are in the SBP compared to 26.9 percent nationally. Thus, while the CLOC and Commodity SFA's in the sample are closely matched for purposes of making comparisons for this evaluation, they are not representative of the nation in all respects. The 25 Commodity SFA's selected for participation in this evaluation were matched to the 25 CLOC SFA's that were already in place. Potential control sites were selected in each of the 22 States in which the CLOC SFA's are located. The potential control sites were selected on the basis of enrollment, number of schools, percent of students nonwhite in 1981, percent of students below the poverty line in 1981, instructional expenditures per pupil and geographic proximity. A list of 30 SFA's was selected for each state using a disjoint cluster analysis technique. From this list and with the advice of the Child Nutrition Director in each state, the final sample was selected. In one case it was necessary to select an SFA from a neighboring State (Montgomery County Schools, Maryland) due to the large enrollment of the CLOC SFA (Fairfax County Public Schools, Virginia) against which it was matched. The SFA's comprising the sample are listed in Table III-l. The adjacent CLOC and Commodity SFA's on each line are paired for purposes of the evaluation. Though an overall sample size of 50 is used in the evaluation, most of the comparisons are of the two groups of 25 SFA's. This size was dictated by the number of SFA's operating under CLOC in SY 1990/91. The adequacy of a sample size is determined by a combination of (a) the degree of variability of the outcome measure, (b) the magnitude of difference that is to be detected, and (c) the desired level of significance. In the earlier evaluation and using Cohen's power tables, the authors concluded that with a sample of about 30, a desired power level of 75-85 percent, and a 5 percent level of significance, differences of less than 0.6 standard deviations were hard to detect. V With a sample only slightly smaller than this, the present evaluation can detect differences of approximately the same magnitude. While these are relatively large differences, compared to those that could be detected with a sample four or five times larger, when observed they will be differences of consequence. B. Data Collection A variety of data was required for this evaluation. The most demanding data need, by far, was that required to compare the farm-equivalent weight and value of individual commodities removed from the market by CLOC and Commodity SFA's. This necessitated the collection of a very large volume of detailed acquisition and inventory records for the entire 1990/91 school year from each of the participating SFA's. For purposes of evaluating other differences between the CLOC and Commodity systems that might affect evaluation results, participating SFA's were asked to provide descriptive information for their operations for SY's 1989/90 and 1990/91. To assist in evaluating the impact on other y St. Pierre, Robert G., et.al., Evaluation of Alternatives to Commodity Dontaion in the National School Lunch Program: Final Report. Abt Asssociates, Cambridge, Mass. Prepared for FNS, USDA, 1986, p.70. 111-2 <90 ffl Table ffl-1 School Food Authorities Participating in the Commodity Letter of Credit Modification Evaluation State Arizona Arkansas California Colorado Connecticut Florida Idaho Illinois Iowa Maine Michigan Minnesota Pennsylvania South Carolina South Dakota Tennessee Vermont Virginia Washington West Virginia Wisconsin Wyoming CLOC Site Flowing Wells Schools Green Forest School District Gilroy Unified School District Weld County School District Windsor Public Schools Hernando County School District Fruitland Public Schools Community S.D. #4 (Paris) Parkersburg Community S.D. Portland Public Schools Troy School District Gonvick-Trail Community S.D. Indiana Area School District Lexington County S.D. #3 Elk Point School District Loudon County School District Lyndon Town School District Alleghany County School District Fairfax County Public Schools Longview School District #122 Shoreline Public Schools #412 Pendleton County School District Merrill School District River Falls School District Washakie Community S.D. #2 Commodity Site Creighton School District Atkins Public School District Hueneme School District St. Vrain Valley RE D Vernon School District Martin County School District Buhl Joint School District Bradley Bourbonnais S.D. 307 Beaman Conrad Liscomb S.D. Lewiston School District Rochester Community S.D. Herman-Norcross S.D. 264 Grove City Area S.D. Saluda County S.D. #1 Sioux Valley School District McMinn County S.D. Swanton School District Floyd County School District Montgomery Cnty Schls (MD) Kelso School District #458 Franklin Pierce S.D. #402 Gilmer County School District S.D. of Menomonie Area Monroe School District Fremont County S.D. #2 commodity donation programs of adopting CLOC nationwide, the 50 State Distributing Agencies that coordinate commodity donations through the NSLP in their respective States were each asked to provide information on their distribution systems. Finally, limited information on the domestic content of CLOC purchased foods was collected through on-site visits to some of the CLOC SFA's. The data collected through these various efforts formed the basis for this evaluation. Because the validity of the evaluation results is directly m-3 Q?/ dependent on the accuracy and reliability of the data that underpin them, a great deal of attention was devoted to the quality of these data. A more detailed description of each of the major data elements follows. Food Acquisition and Inventory Pata To compare the volume and value of foods removed from the market by the CLOC SFA's versus that removed by the Commodity SFA's, it was necessary to collect information on all foods entering these school systems throughout the year under study. To help make the data collection effort more manageable and less costly, the scope of the inquiry was limited to foods derived from the following thirteen commodities: beef apples green beans pork peaches green peas chicken apricots potatoes turkey pears sweet corn tomatoes CLOC SFA's continued to receive CLOC's for the full range of commodities that were donated in SY 1990/91. Information on individual CLOC purchases for the full range of commodities was obtained from Virginia Polytechnic Institute and State University, the independent contractor that administers the CLOC program. However, acquisition data were collected directly from SFA's for only these thirteen commodities. Collectively, these thirteen commodities accounted for approximately 75 percent of the total value of entitlement commodities distributed through the NSLP in SY 1990/91. In total, approximately 350 unique food products containing these thirteen commodities were acquired or used by these SFA's over this period. Food used by these school feeding programs over the 12-month period of study originated from one of three possible sources -- commercial purchases, USDA commodity donations or beginning-of-the-year inventories. Foods acquired from these sources are either used during the year or held in end-of-the-year inventories, assuming there is no waste or spoilage. Since the two measures of relevance to this evaluation are annual acquisition and annual usage, information was required for: commercial purchases, USDA commodity donations, beginning inventories, and ending inventories. Because volume data for food transactions at the SFA level are reported in a variety of forms, including case sizes, it was necessary to convert all units of volume measure to pounds. This was accomplished through use of a conversion system based on standard industry conversion factors published in USDA's Conversion Factors and Weights and Measures (Statistical Bulletin No. 616). To estimate the farm-equivalent weight of the commodities contained in these foods it was further necessary to apply an additional set of "commodity conversion" factors. These factors translate the weight of the processed product as it is acquired by the SFA into the raw commodity weight, adjusting for changes in the weight of the commodity as it is slaughtered, pared, processed, transported, deboned, cooked, combined with other ingredients, etc. in moving from farm to SFA. Commodity m-4 Q3 conversion factors for this evaluation were provided by FNS from a combination ofrecently calculated factors for newer food products and the FNS archive of conversion factors derived from earlier studies of farm impact.^ Different data elements were collected, depending on whether it was an inventory, donation, or purchase item. The data elements collected for each were as follows: Inventory Product name Product description Unit size Number of units Value If USDA donation Donations Commodity name Unit size Number of units received Month cf delivery Weight USDA value Purchases Product name Product description Unit size Number of units Unit cost Total cost Delivery month Rebate or discount Weight Evaluation staff consulted with each SFA participating in the evaluation to determine how the necessary data could be collected with the least burden to them. To help defray the out-of-pocket costs associated with copying records and mailing them to the data collection office, an administrative allowance of $250 to $1,000, based on the number of reimbursable meals served in October 1990, was paid to each participating SFA. Inventory data were provided by most SFA's in the same form that they recorded it for other purposes. Delivery slips were the most frequent form in which information on commodity donations was provided, while food purchase information was generally supplied in the form of copies of invoices obtained from vendors at the time of their deliveries. These data were assembled, copied, and mailed to the evaluation data collection office, usually on a monthly basis once the project was underway. On receipt of the data, evaluation staff logged-in the data; examined it for legibility, completeness, and possible duplication; transcribed the data to a standardized transcription form; aggregated the transcribed entries as appropriate (e.g., combined multiple purchases of the same food item within a given month); assigned a numeric foodcode to each entry; entered the transcribed data in a micro computer; completed a series of edit checks on the entered data; and printed a hard copy of the entered data for further editing. SFA Operational Data General descriptive information was required for each of the SFA's participating in the evaluation for purposes of developing a common unit of measure that could be used in quantifying and comparing the overall level of activity of each SFA. Development of this unit of measure, the lunch equivalent (LEQ), is described later in this Chapter. These data V St. Pierre, loc. cit. m-5 =23 J were also needed to specify a regression model (also described later in this chapter) that was used to adjust observed mean values for such non-treatment differences as school district size, wage rates, and type of distribution system used. In addition to the operationaldata that were collected, SFA managers were asked to report significant changes that had occurred within their systems during the study period, and CLOC SFA managers were asked for their opinions regarding the effect of the CLOC modifications on their operations. These data were collected by two nearly identical mail surveys, one collecting data for SY 1989/90 and the other for SY 1990/91. The survey instrument was pre-tested with three of the participating SFA's in December 1990. The SY 1989/90 survey was mailed to the 50 participating SFA's on April 10, 1991; the SY 1990/91 survey was mailed May 16, 1991. The response rate for both surveys was 100 percent. However, substantial telephone contact between evaluation staff and the respondents was required to achieve clarification of responses and to ensure that questions were interpreted consistently. Having to respond to two significant data requests in such a short period of time, in addition to providing copies of food acquisition and inventory records at the same time, was a burden on many of the SFA's participating in the evaluation. Adding to this difficulty was the request for annual data for the school year just ending, data that were not yet available for many SFA's at the time they received the surveys. It is to the credit of these SFA managers that this large volume of infoimation was provided on such an expeditious basis. Five identical types of information were collected on both surveys. They were: • General information regarding program responsibility and any major changes that had occurred during the school year. • Information on enrollment, number of serving days, number of schools, attendance, and participation in school feeding programs. • Information relating to the meal production and delivery systems used. • A detailed listing of staff, hours worked, and wages. • Basic information about the SFA's expenses and income for the school year. In addition to this information, the second survey collected SFA managers' opinions on program administration. For CLOC SFA's, the questions related to the effect of the CLOC modifications on different dimensions of their operations. State Distribution Systems One of the research questions to be addressed through this evaluation concerns the effect on other commodity donation programs of adopting CLOC nationwide. To answer this question, it is necessary to know what types of State distribution systems are currently ffi-6 SN in use. An earlier study described these systems as they were operated in 1985/86,^ but more recent information was not available. Since legislation has been adopted since 1985/86 requiring States to consider changes in their distribution systems and because some States had made changes in their systems, it was decided that more current information was desirable. This information was collected through use of a 2-page mail survey sent to he SO SDA's responsible for coordinating commodity donations through the NSLP. The surveys were mailed April 18, 1991. Completed surveys were received from all respondents. To reduce respondent burden and help insure consistency with results of the 1985/86 study, each SDA received a unique instrument chat displayed the information previously provided for their State and asked them to up-date the information to 1989/90. The year preceding the year of study was used since complete data for 1990/91 would not have been available for all SDA's at the time they were preparing their Domestic Content Data Another of the research questions to be investigated in this evaluation required information on food imports and on the country of origin of commodities contained in CLOC-purchased foods. As a first step, data on food imports were obtained from the computerized trade data system maintained by the Foreign Agriculture Service (FAS) of the U.S. Department of Agr~ulture. These data, collected by the U.S. Department of Commerce, are published monthly by FAS. For purposes of this analysis, data on 51 food product categories were tabulated for the past five years including U.S. imports measured both by volume and value by country of origin. As a basis of determining the relative importance of imports of individual food products, the volume of imports was then matched with comparable data for domestic production collected from a variety of trade and government sources including the Almanac of the Canning. Freezing, and Preserving Industries, the American Meat Institute, Milling and Baking News, and the USDA's Economic Research Service. This review identified eight foods that were imported into the U.S. in sufficient volume during 1988/89 to warrant examination in this study: apple juice, frozen boneless beef, canned tuna, tomato products, canned and fresh pears and peaches, and canned and frozen sweet corn and green peas. Based on consultation with Federal agency and private industry sources, a procedure was designed using can codes for collecting information on the country of origin of processed foods and a form was developed for use in collecting the required information from SFA warehouses. The form was pre-tested at the warehouse of a participating CLOC SFA in January 1991. In April 1991, representatives of FNS Regional Offices conducted on-site visits at an additional 18 CLOC SFA's to collect information on any of the foods described above held in storage by the SFA. The information collected for each food item included: food item V Abel, Daft & Earley, A Study of the State Commodity Distribution Systems, a report prepared for FNS, USDA, March 1988. IH-7 33 descriptor, unit/case size, label information, can code, manufacturer name and address, and distributor name, address, contact, and telephone number. C. Analytic Techniques Most of the analyses in this evaluation are designed around comparisons of two groups. Answers to some questions require a comparison of the CLOC SFA's and the Commodity SFA's; answers to other questions require a comparison of the CLOC SFA's prior to the modifications with the same SFA's following the modifications. Almost all of the comparisons across time are between SY 1989/90 and SY 1990/91. For purposes of making comparisons, mean values are calculated for each group (e.g. for CLOC SFA's) with the measure for each SFA within its respective group assigned an equal weight. Statistically significant differences between mean values are determined on the basis of calculated ^statistics. When comparing CLOC and Commodity sites, the t-test for individual samples was used. When comparing the same CLOC sites across time, the t-test for paired samples was used. For purposes of interpreting results, two-tailed tests are used and differences significant at confidence levels of .10, .05 and .01 are identified. Because the paired SFA's have been closely matched on the basis of such characteristics as size of enrollment, number of schools in the system, population density, and incidence of poverty, the effect of exogenous factors on the comparisons has been reduced. Though a sample of 50 SFA's is small relative to the national total of over 15,000, it is quite adequate for determining significant differences between the two treatments. Enlarging the sample would add very little power to the use of the t-test. For example, in going from a sample of 50 to one that is infinitely large, a two-tailed t-test at the .01 confidence level only falls from 2.682 to 2.576. Regression Adjusted Means Despite the careful matching of the CLOC and Commodity samples, there remains the possibility that other differences between the two groups could influence the outcome of the comparisons. To determine whether these differences affected the outcome, a second comparison was made using mean values calculated on the basis of a regression model developed for this purpose. Variables included in the regression are as follows: Size. Size is represented by the total number of potential participants in the NSLP for each school system as of October 1, 1990. This information was collected through the SFA mail survey of their SY 1990/91 operations. The number of potential participants is defined as the total number of enrolled students who have access to the program. It differs from total enrollment to the extent, for example, that there are schools in the system that do not offer the NSLP or there are children in kindergarten who do not have an opportunity to eat lunch due to Vi-day sessions. The number of potential participants in the sample ranged from a low of only 124 to a high of over 1,900. Free and reduced price lunches. This measure is represented by the total number of students approved for free and reduced price lunches as a percent of the number of m-8 6£o potential participants in the NSLP, both as of October 1, 1990. This variable represents a combination of influences. In part it is a measure of the income level of families with students enrolled in the school districts and, therefore, is an indicator of the economic well-being of the community. It also reflects the price structure charged students for the meals they are served as well as the relative importance of the NSLP to the SFA's meal program. The higher the share receiving free and reduced price meals, the higher the degree of subsidization and the less important paid meal price as a determinant of participation. Among SFA's in the sample, the share of free and reduced price eligibility ranges from 2 percent to 73 percent. Participation rate. This measure is calculated by dividing the total number of reimbursable lunches served during SY 1990/91 by the product of the number of potential lunch participants (adjusted by an attendance factor) and the number of serving days. Where the number of serving days varied within a given SFA, say, between the elementary and secondary schools, an adjustment was made based on relative enrollments. By its nature, this is a partial measure of participation in an SFA's overall feeding program since it is limited to those who eat reimbursable lunches and excludes those who eat a la carte meals. Not surprisingly, there is a relatively large (.48) positive correlation between participation rate and free and reduced prices (because of the first factor) and a relatively large (-.58) negative correlation between participation rate and a la carte sales (for the second reason cited above). Calculated participation rates among the sample SFA's- ranged from a low of 32 percent to a high of 99 percent. On-site food preparation. This variable represents possible differences in organization and methods of food preparation. An "on-site kitchen" is defined as a cafeteria where all meals served are prepared in the kitchen of that cafeteria. The variable was calculated by dividing the number of "on-site kitchens" by the total number of kitchens for each SFA. The prevalence of independent kitchens (as opposed to the use of central kitchens) could have an effect on the costs of operation as well as on the types, forms, and quantities of food used. The share of on-site kitchens in the sample ranged from 0 percent to 100 percent with 22 of the SO SFA's using on-site kitchens exclusively. Commercial Distribution. This variable measures the SFA's use of commercial distributors to deliver USDA-donated commodities. It is represented by an indicator variable that is equal to "1" if commercial distributors served the SFA and "0" otherwise. It is included on the premise that those SFA's that receive their commodity donations through commercial distributors might enjoy greater flexibility in the timing of deliveries and might experience lower costs of storage, if not lower overall operating costs. Of the 50 sample sites, 9 Commodity SFA's reported having donated commodities delivered by commercial distributors and 8 CLOC SFA's had bonus commodities delivered in a similar fashion. A la carte food sales. This variable represents that portion of food sales that falls outside the range of reimbursable meals because it doesn't meet the criteria of a reimbursable meal. Some SFA's provide extensive a la carte options to their students, while others provide no a la carte options at all. Since a la carte sales tend to be concentrated within certain food categories (e.g. hamburgers, french fries, pizza), a high share of a la carte sales could influence the rate of usage of certain foods. A la carte food sales (and other food service income including income from adult meal sales and catering sales) are measured as a ffl-9 Q-i percent of total SFA income. This share ranges from 5 percent to 61 percent among SFA's in the evaluation. Labor cost. Since labor is one of the two most important expenditure items in SFA budgets (purchases of food is the other), the rate at which labor is paid can have an important influence on how the SFA is organized, how much food preparation occurs within the SFA, and the form in which foods are purchased. For purpose of the regression model, labor cost is measured as total labor expense, including fringe benefits, reported by the SFA for SY 1990/91 divided by the total number of hours worked in SY 1990/91 by all food service staff. In those cases where SFA staff were paid out of other budgets for work performed in food service, the data were adjusted to reflect the actual hours worked and associated costs. Treatment. Treatment is an indicator variable with "1" representing CLOC SFA's and "0" representing Commodity SFA's. The very low correlation between this variable and each of the others is indicative of the successful pairing of treatment and control sites. A correlation matrix for the independent variables included in the regression model appears in Table III-2. The model was used to regress the measures that were being compared for CLOC and Commodity SFA's (e.g. the farm-equivalent weight of beef usage) on the eight independent variables described above. The resulting set of coefficients was used to derive adjusted mean values for each of the measures that were being compared. For the CLOC SFA's, this was done by plugging the mean values for the CLOC SFA's for each of the independent variables into the regression. The same procedure was used to derive an adjusted mean value for the Commodity SFA's. The statistical significance of any difference in these adjusted means was determined by the t-statistic for the treatment variable in the regression. Food Acquisition/Food Use Several of the questions around which this evaluation is designed are framed in terms of the removal of commodities from the market during a given school year. This removal occurs in one of two ways. Either foods are purchased directly by SFA's from commercial sources or they are purchased by USDA and distributed to participating schools as commodity donations. SFA purchases and commodity donations, collectively referred to as "food acquisitions," are captured through the data collection procedures described above. Yet, food acquisitions are not the same as "food use." Usage is a function of the combined effects of acquisition and net changes in inventory. Large buildups or drawdowns of inventory can result in substantial differences in acquisition and usage. And while the perishability and costs of storage of food generally dictate small inventories for food service operations, difference of varying degree exist among different types of food. Where appropriate in this evaluation, results will be calculated using both measures. Those based on food acquisitions should be considered partial and indicative of short-term impacts. Those based on food usage provide a more complete picture, though one that will take a slightly longer period of time to become evident. m-io 3^ Table ffl-2 Correlation Coefficients of Variables used in the Regression Model Use of Ala Free and Participation On-site Commercial carte Labor Variables Size reduced price rate preparatior1 dist. sales cost Treatment Size 1.00 -.24 -.35 -.22 -.08 .36 .38 .02 Free and reduced price 1.00 .48 .12 .21 .53 .02 -.04 Participation rate 1.00 .41 .08 -.58 -.41 -.04 On-site preparation 1.00 .03 -.34 -.45 .01 Use of commercial distr. 1.00 -.02 -.10 -.04 A la carte sales 1.00 .30 .04 Labor cost 1.00 -.09 Treatment 1.00 o?9 D. Special Measurement Issues and Problems Lunch Equivalents To compare different school feeding programs, it is necessary to find a common yardstick against which they can all be measured. It is not enough to say that School A uses twice as many pounds of ground beef as School B, for example, given that the schools probably differ in many respects including the number of students they serve, the rate of school lunch participation, etc. In other words, how can these different levels of usage be put on a uniform basis so that meaningful comparisons can be made between results for the CLOC and Commodity samples and so that inferences can be drawn regarding the larger populations these schools represent? One way to achieve comparability is to adjust for differences in the level of output. In the case of school feeding programs, there are three major types of products: (1) lunches that meet the USDA meal pattern requirements and are eligible for reimbursement, (2) breakfasts that meet the USDA meal pattern requirements and are eligible for reimbursement, and (3) other food sales, including a la carte sales to students and adult meals to school staff. Definitions of the first two categories are prescribed by USDA in terms of minimum quantities of specified food types that are to be served. In filing for Federal reimbursements, participating schools are required to report the number of reimbursable meals served every month. The principal measurement issue for these categories, therefore, is determining their comparability. Is one lunch equal to two breakfasts, for example, or to 1.28 breakfasts or to some other ratio? Once that is decided, weights can be assigned and a measure of output can be determined. It is the third type of product, a la carte and other food sales, that presents the greatest measurement challenge. Since individual sales transactions in this category can include everything from an ice cream bar or a single order of french fries to a full (non-reimbursable) meal, the number of transactions is not itself a meaningful measure. The only practical way to measure the volume of sales in this category is on the basis of the dollar value of sales. The issue, therefore, becomes one of translating the dollar value of other food sales into lunches and breakfasts or vice versa. In the evaluation of the Commodity Donation Demonstration that was conducted during 1982-86, Abt Associates, Inc. (AAI) defined a lunch equivalent (LEQ) using a simple quadratic model based on two price indexes, one for food and one for labor. An LEQ, by this methodology, was the weighted sum of the number of lunches and breakfasts and the revenue from other food services produced by an SFA.-^ With this model, AAI estimated • St. Pierre, Robert G. et. al., Evaluation of Alternatives to Commodity Donation in the National School Lunch Program: Final Report. Vol 6, Abt Associates, 1986, Appendix C. Ill-12 LEQ's for each SFA in the study sample. For average values of the food and labor price indices, a breakfast was found to approximate 0.78 lunches and each dollar of revenue from other food sources was found to approximate 0.97 lunches. Mere recendy, AAI, as part of the FNS sponsored Child Nutrition Program Operations Study, has developed a new procedure for estimating LEQ's. Under the new estimating procedure, an LEQ is defined as "the number of lunches necessary to generate the same expected cost for a given SFA as it actually did incur to produce its output of breakfast, lunches, and other items." ^ Using cost information from 1,180 observations, AAI estimated three cost functions. Two of these functions were applied to data for the 50 SFA's included in this evaluation. In these functions, total SFA cost is estimated on the basis of the number of reimbursable lunches, the number of reimbursable breakfasts, and the dollar value of other food sales. Using this estimated cost and setting the number of breakfasts and the dollar value if other food sales equal to zero, the equation is then solved for LEQ. Results from applying the more complex of the two functions to the data for this evaluation was unsatisfactory to the extent that results for certain SFA's, particularly for small SFA's that served breakfasts, deviated significantly from expected levels. Also, since cost data had been collected from SFA's participating in the evaluation, it was possible to compare estimated cost to reported cost and they did not match-up well either. The simpler of the two functions developed by AAI proved to be a far more accurate predictor of reported cost. The resulting estimates of LEQ were also found to be much closer to the expected values than those estimated by the first function. As a result, this function was used in estimating LEQ's for each of the participating SFA's based on data provided for SY 1990/91. The function is as follows: Total SFA costs = -42,428 + 1.91 (number of lunches) + 1.97 (number of breakfasts) + 0.46 (value of other food sales) Valuing Donated Commodities In comparing the CLOC and Commodity systems, the value of the benefits received by the participating SFA's is one of the measures compared. For CLOC SFA's, valuation is straightforward for entitlement commodities since the benefit they receive is in monetary form. For the Commodity SFA's, however, the valuation of benefits is more complicated V St. Pierre, Robert G., et. al, Child Nutrition Prop-am Operations Study: First Year Report. Abt Associates, August 1991, Appendix I. ffl-13 0 I since they receive their benefits in-kind. USDA assigns a value to its donations that reflects the acquisition price USDA has paid and the cost of transporting donations to State delivery points. For certain purposes this is a satisfactory measure. For example, one of the researchquestions requires a comparison of Commodity SFA entitlement to the value of the donated commodities they each received. Because the volume of donations Commodity SFA's are eligible to receive under their entitlement is based on the use of USDA assigned values, this valuation is appropriate for answering this question. For answering other questions, however, it is less appropriate. For example, another research question compares the value of CLOC issuances per reimbursable lunch to the value of commodity donations per reimbursable lunch. If donated commodities are valued on the basis of USDA values, it is anticipated that the resulting measures for CLOC and Commodity SFA's will be approximately equal. That is, since the entitlement levels for SFA's in both systems is based on the number of reimbursable lunches they serve, if the schools in both systems are receiving the level of benefits to which they are entitled, the relationship will be roughly equal. Yet, if these donations are assigned comparable commercial values, it is expected that they will exceed the value of CLOC issuances per reimbursable lunch. This occurs because when compared to a commercial sales price paid by an SFA for food delivered to its door, the USDA value is understated. It is understated because it excludes costs associated with the operation of a procurement, scheduling and order-taking system as well as costs associated with the storage and transportation of commodities from their receipt at state delivery points to delivery at SFA's. Costs of inspection and auditing are also excluded from this value. In evaluating the earlier demonstration pi oject, Abt researchers concluded that the difference in entitlement subsidy attributable to those functions was worth approximately 3<t per LEQ (relative to a commodity entitlement of 12? per NSLP lunch in SY 1984-85). Given these differences and the different interpretations that accompany them, USDA donated commodities will be valued on two bases in this evaluation. For certain purposes the USDA assigned values will be used. For other purposes commercial values of comparable products will be used. Both approaches will be adopted in answering some questions so that any differences between the two can be evaluated. For other questions, the most appropriate basis of evaluation will be used. A single national value was used for valuing each donated item. If USDA purchased the item more than once during the school year, the value of the last purchase made was used. A comparison of purchase prices paid for meat and poultry products in SY 1990/91 indicated that, on average, the price used as a result of following this methodology was 98.7 percent of the a
Object Description
Title | Commodity letter of credit modification demonstration evaluation final report |
Date | 1992;1995 |
Creator (individual) | Daft, Lynn M. |
Contributors (individual) | Westfall, Donald W. |
Contributors (group) | United States Food and Nutrition Service Office of Analysis and Evaluation.;Abel, Daft & Early. |
Subject headings | National school lunch program;Food relief--United States |
Type | Text |
Format | Pamphlets |
Physical description | 1 v. (various pagings) :ill. ;28 cm. |
Publisher | Alexandria, Va. : U.S. Dept. of Agriculture, Food and Nutrition Service, Office of Analysis and Evaluation, |
Language | en |
Contributing institution | Martha Blakeney Hodges Special Collections and University Archives, UNCG University Libraries |
Source collection | Government Documents Collection (UNCG University Libraries) |
Rights statement | http://rightsstatements.org/vocab/NoC-US/1.0/ |
Additional rights information | NO COPYRIGHT - UNITED STATES. This item has been determined to be free of copyright restrictions in the United States. The user is responsible for determining actual copyright status for any reuse of the material. |
SUDOC number | A 98.2:C 73 |
Digital publisher | The University of North Carolina at Greensboro, University Libraries, PO Box 26170, Greensboro NC 27402-6170, 336.334.5304 |
OCLC number | 903978300 |
Page/Item Description
Title | Part 1 |
Full-text |
COMPLETED
United Stales
Department of
Agriculture
Food and
NubWon
Service
Office of
Analysis and
Evaluation
M7V-A-M Aft.2:C73
Commodity Letter of
Credit Modification
Demonstration
Evaluation
Final Report
tf
T
ABEL, DAFT & EARLEY
1410 KING STREET. ALEXANDRIA, VA 22314
TEL: (703) 739-9090 FAX: (703) 739-9098
COMMODITY LETTER OF CREDIT MODIFICATION
DEMONSTRATION EVALUATION
FINAL REPORT
by
Lynn M. Daft
and
Donald W. Westfall
Abel, Daft & Earley
Alexandria, Virginia
submitted to:
John R. Endahl
Food and Nutrition Service
U.S. Department of Agriculture
3101 Park Center Drive, Room 204
Alexandria, Virginia 22302
March 1992
T
EXECUTIVE SUMMARY
Introduction
The National School Lunch Program (NSLP) provides Federal assistance to the
feeding programs of the nation's elementary and secondary schools. This assistance is
provided in two forms: cash and donated commodities. For the past ten years, beginning
in School Year (SY) 1982/83, 25 school districts throughout the nation have participated
in a demonstration of an alternative to the existing commodity donation component of the
NSLP. Under this program, school districts are authorized through letters of credit to make
their own purchases of specified foods in place of receiving donated commodities purchased
by USDA. The commodity letters of credit (CLOCs) are used to purchase foods from local
commercial sources.
In response to criticisms of some elements of CLOC, the program was modified
in SY 1990/91. This report describes the results of an evaluation of the effects of these
modifications. The evaluation began in September 1990 and was conducted for the Food
and Nutrition Service of the U.S. Department of Agriculture by Abel, Daft & Earley of
Alexandria, Virginia.
Background
The CLOC alternative arose out of a desire by some School Food Authorities
(SFA's) for changes in certain features of the commodity donation program as it was
operated in the 1970's. For example, some commodities were received in forms that could
not be conveniently used in preparing meals (e.g., whole turkeys, large blocks of frozen
ground beef, and 100 pound bags of flour). Some of the foods were of low acceptability to
students and therefore difficult to use. Some schools lacked control over the size and timing
of commodity deliveries. As a result, donated commodities often had to be stored for as
long as several months. For these and other reasons, there had been a continuing interest
on the part of some schools in alternatives to commodity donation. This interest ultimately
led to a Congressional directive, as part of the 1981 Agricultural Appropriations Act, to
conduct a 3-year study of two alternatives to the existing commodity donation program --
an all cash approach and CLOC.
The study got underway in SY 1982/83 with 96 school systems from 29 States
assigned to one of the three options under study: commodity donation, cash, and CLOC.
The study ran through SY 1984/85. The cash school systems in the study were allocated
cash equivalent in amount to what the USDA would have used in buying commodities for
donation to them. The SFA's, were free to use this cash in support of their school feeding
programs in any way they wanted.
II
CLOC SFA's also received cash equivalent to the value of the donated
commodities that they would have received, but the expenditure of this cash was restricted
to foods containing those commodities that USDA was buying for donation through the
NSLP. In this way, CLOC funds were used to provide market support for the same
commodities USDA was supporting through its purchases. CLOC funds were allocated
among commodities in the same general proportion as USDA purchases. Authorization for
the use of funds (through the issuance of CLOC's) was timed to coincide with USDA's
purchases, as were deliveries of these purchases. To provide SFA's with maximum flexibility
in their use of these funds, partial credit was granted for the purchase of a wide range of
foods containing the particular commodity. The amount of credit was dependent on the
share of product weight accounted for by the commodity.
Final results of the evaluation of this initial demonstration of CLOC were
published in 1986.-^ On the one hand, they demonstrated that CLOC was administratively
feasible at the local level and that most of the SFA's participating in CLOC were pleased
over its effects on their individual feeding programs. On the other hand, CLOC schools
removed less of some commodities (e.g. beef, chicken, pears, and sweet potatoes) from the
market. Related to this, CLOC schools bought more highly processed foods, in effect
spending a larger share of their Federal assistance on processing and a smaller share on
commodities. Results of the cash option were similar to those for CLOC, though the
reduced level of commodity support was even more pronounced, at least for certain
commodities.
At the conclusion of this evaluation, the USDA recommended that CLOC be
discontinued. However, SFA participants in CLOC successfully appealed to the Congress
for its continuation as a pilot project. The program has continued to operate on this basis
ever since.
The commodity donation program has been improved in a number of respects over
the past 5 to 10 years. This includes reduction in the fat and salt content of foods
distributed , improved labeling and product identification, inc.eased product varieties,
improved scheduling of deliveries, improved packaging, the adoption of unitized deliveries,
and greater attention to nutrient analysis.
In 1989, the Conference Report accompanying the Child Nutrition and WIC
Amendments Act authorized the Secretary of Agriculture to design and test modifications
of CLOC that would address shortcomings of the program identified in the earlier
evaluation. Later in the year, the Secretary directed the Food and Nutrition Service to
develop modifications of CLOC based on consultation with the affected interests and to
conduct an evaluation of the effects of these modifications.
* St. Pierre, Robert G., et. al. Evaluation of Alternatives to Commodity Donation
National School Lunch Program: Final Report. Volume o, 1986.
u
Ml
Commodity Letter of Credit
Modifications
The CLOC program was modified in a number of important respects, effective
at the beginning of SY 1990/91. The program was modified for all 25 school systems that
were still participating in CLOC. The most important of these changes were the following:
• Specifying the commodity to be purchased. The USDA now determines on a
case-by-case basis whether CLOC's are to be issued for generic commodities (e.g.
"pears"), as they were in the past, or for more specific commodities (e.g. "fresh
winter pears").
• Full and partial credit. Authorizations for the use of CLOC funds are now based
on achieving a farm-equivalent weight M of the targeted commodity per dollar
spent comparable to that of USDA purchases. This has the effect of narrowing
the range of products for which credit is authorized and reducing the amount of
partial credit allowed for many foods.
• Timing of purchases and deliveries. The time periods allowed for CLOC
purchases and deliveries have generally been accelerated and shortened to better
match the schedule of the USDA commodity program. For CLOC purchases of
fruits and vegetables the reduction was especially large.
• Trades, rollovers, and substitutions. The rollover of small balances remaining in
CLOC accounts and limited substitution among commodities, both of which were
permitted under the original CLOC concept, have been discontinued. And, the
trading of letters of credit of one commodity for those of another commodity,
among participating CLOC SFA's are no longer required to be of equal dollar
amounts.
• Domestic origin. SFA's and their distributors are required to certify that all foods
purchased with CLOC funds are of domestic origin.
Evaluation Objectives and Methodology
There are two central issues to be addressed in this evaluation, both related to the
CLOC modifications adopted in SY 1990/91. They are:
(1) The effect of these modifications on the performance of the CLOC program
relative to the performance of the existing Commodity Donation Program.
y Farm - equivalent weight is the weight of a raw agricultural commodity at the farm gate
that is required to produce a given unit of a food product containing that commodity.
For example, it takes approximately 1.25 pounds of raw apples coming out of the
orchard to produce 1 pound of canned applesauce.
• ■ • m
Y
(2) The effect of these modifications on the performance of the CLOC program now
relative to its performance prior to the modifications.
The Congressional authorization for these modifications made clear that it was looking for
modifications that would overcome criticisms of the earlier program without undermining
the CLOC concept's central strengths.
These general issues were first translated into a list of 14 research objectives and then
into a list of about 20 researchable questions. The list includes such questions as:
Do CLOC and Commodity SFA's remove the same weight of each commodity
per lunch served?
Do CLOC SFA's buy more highly processed foods?
To what extent are CLOC's not used or traded and how does this compare to the
past?
How do the level of benefits received by CLOC and Commodity SFA's compare?
How does the timing of CLOC issuances, purchases, and deliveries compare with
the timing of USDA purchases and deliveries?
To what extent are CLOC's used to purchase foods that are not of domestic
origin?
What effect did the CLOC modifications have on labor use and costs and or rates
of participation in the NSLP?
How do CLOC SFA managers view the modifications?
Of the 32 CLOC SFA's that participated in the original demonstration, 25 remain in the
program. Because participation in the CLOC program is voluntary, those SFA's that remain
in it apparently prefer CLOC to the existing commodity donation program. Furthermore,
having participated in CLOC for nearly 10 years, these SFA's have had time to learn CLOC
procedures and were therefore reasonably well prepared to adjust to the modifications.
For purposes of this evaluation, 2Z additional SFA's were selected as matched control
sites. They were matched to the CLOC SFA's on the basis of a number of criteria including
size of enrollment, number of schools, percent nonwhite, and percent below the poverty line.
With one exception, control sites were matched with CLOC sites from the same State. The
one exception was made necessary by the large enrollment of the SFA against which it was
matched.
Each of 50 SFA's participating in the evaluation provided several types of data, some
of it in highly detailed form. To determine the volume and value of food acquisitions and
IV
<
use, each SFA provided records describing food purchases, commodity donation deliveries,
and beginning and ending inventories for all foods containing any of 13 commodities for SY
1990/91. y The evaluation was restricted to 13 commodities to help make data collection
more manageable and less costly. For the most part, this information was provided in the
form of copies of invoices, delivery sups for donated commodities, and ending inventories.
These data were periodically mailed to a data collection office where they were transcribed,
entered into micro computers, and edited.
Participating SFA's also provided operating data for their operations for SY's 1989/90
and 1990/91 through two mail surveys. A 100 percent response rate was achieved for both
surveys. Information on the State commodity distribution systems was collected through a
mail survey of the 50 State Distributing Agencies (SDA's) that distribute commodities
through NSLP. All 50 agencies provided the information that was requested. Finally, limited
information on the domestic content of foods purchased with CLOC funds was collected on-site
by FNS Regional Office Staff from 18 CLOC SFA's.
Most of the research questions have been analyzed through a comparison of mean
values. For some questions these are comparisons of modified CLOC and Commodity
SFA's; for other questions they are comparisons of CLOC SFA's before and after the
modifications were adopted. Statistically significant differences have been determined using
the two-sample t-test and the paired t-test, as appropriate. In answering some research
questions, values have been adjusted through use of a regression model designed to
eliminate the influence of a variety of exogenous factors such as enrollment size and wage
rates.
Evaluation Findings
The most significant findings of the evaluation are those related to the two central
issues described earlier, i.e., how have the CLOC modifications affected (a) CLOC
performance relative to Commodity performance and (b) the advantages of CLOC as it has
operated in the past? Overall, the findings indicate that a number of the major differences
in the performance of CLOC and Commodity SFA's found in the previous study ai.d that
were the focus of the CLOC modifications have been narrowed or eliminated. Furthermore,
the principal attractions of CLOC were found to have been largely unaffected by these
modifications. The more detailed findings are described below.
Market removals. Almost no significant difference was observed in the farm-equivalent
weight of commodity acquisition or use adjusted by the number of lunch
V The 13 commodities are: beef, pork, chicken, turkey, apples, peaches, apricots, pears,
green beans, green peas, potatoes, sweet corn, and tomatoes.
VI
T
equivalents U (LEQ's), between CLOC and Commodity SFA's. In the case of only one
commodity, pears, was it found that CLOC SFA's used a larger volume per LEQ than
Commodity SFA's. Of all the detailed findings, this is perhaps the most central. It says that
the modified CLOC program provides support for commodity markets over the period of
a school year comparable to that provided by the commodity donation program.
Product composition. Most of the differences in the product composition of
acquisitions by CLOC and Commodity SFA's are related to the form in which commodities
are donated. This indicates that when SFA's are given discretion in their choice of product
forms, they frequently opt for a form other than that in which commodities are donated.
Whole turkey, cut-up chicken, frozen peaches, frozen ground beef, and bulk apple juice are
examples of products that accounted for a significantly larger share of Commodity SFA
purchases.
While there is some evidence that CLOC SFA's acquire more highly processed
foods, the relationship is weak. Only in the case of chicken were Commodity SFA's found
to have acquired products with a higher share of their cost attributable to farm value.
Timing of purchase and delivery. Modifications in CLOC resulted in a substantial
reduction in the length of the purchase period, especially for fruit and vegetable CLOC's.
The average purchase period was reduced from 172 days in SY 1939/90 to 39 days in SY
1990/91. Nearly all CLOC purchases were made within the prescribed time limits. As a
result, in SY 1990/91, CLOC purchases coincided very closely with the timing of USDA
purchases.
Although CLOC procedures allow for a delivery period of at least 3 or 4 months,
most CLOC SFA's arrange for delivery within the purchase period to minimize the
paperwork involved. This has resulted in CLOC SFA's taking delivery of their purchases 3
or 4 months before the corresponding commodity donations reach Commodity SFA's.
Benefit levels. The overall value of the benefits received by CLOC and Commodity
SFA's, adjusted for the number of lunches served, were found to be roughly comparable.
For individual commodities, however, relatively large variations were observed. For
example, Commodity SFA's received over 30 percent more pork than CLOC SFA's on a
dollars per reimbursable meal basis. Most of these differences are due to operational
differences fundamental to the two approaches rather than to any decisions made at the
SFA level.
A comparison of the value of commodity donation receipts by Commodity SFA's
with an estimate of the value of individual commodities to which they were entitled reveaied
y The lunch equivalent measure converts numbers of lunches and breakfasts and the dollar
value of other food sales (e.g. a la carte and adult meals) to a common unit. It is based
on the number of lunches that could have been produced for the same cost incurred in
producing a combination of lunches, breakfasts, and other food sales.
vi
VH
a relatively large variation in the share of entitlement received. This variation occurs both
among commodities and among SFA's. Of the 25 Commodity SFA's, over one-third
received commodities that deviated from their entitlement value by 20 percent or more,
some higher and some lower. These variations in entitlement benefits received by
Commodity SFA's result in part from SFA's refusing all or part of a particular commodity
offered them, SFA's accepting other SFA's refused commodities, limitations on case-splitting
for small SFA's, and variations in the methods SDA's use to calculate total entitlement. The
implication of this finding is that there is a far larger degree of variation in the level and
composition of commodity donation benefits than in CLOC benefits.
CLOC utilization. Of the total value of CLOC's issued in SY's 1989/90 and
1990/91,99.8 percent was spent. The CLOC modification had no appreciable effect on this
share; in fact, the share spent increased marginally in SY 1990/91. CLOC trades among
participating SFA's have not been used extensively, though the level of activity increased
between SY 1989/90 and SY 1990/91 as the share of CLOC value traded rose from 0.2
percent to 0.4 percent. CLOC refusals and rollovers, option.1; that were discontinued under
the modifications, made it possible for CLOC SFA's to substitute among commodities and
to more completely exhaust CLOC funds in SY 1989/90, though neither option was
extensively used.
Labor use. No significant differences between CLOC and Commodity SFA's were
observed regarding changes in either labor hours or labor cost. This is an indication that
the CLOC modifications did not adversely affect labor use. Both measures were adjusted
by the number of lunch equivalents.
Participation rates. The CLOC modifications had no significant effect on rates of
participation in school lunch, school breakfast, or a la carte sales.
Effect of CLOC on other commodity donation programs. Of the 50 State
Distributing Agencies (SDA's) that are responsible for the administration of commodity
donations through the NSLP, ten would have to make major changes in their distribution
systems. On average, these SDA's devote approximately 15 person-years to warehousing
and delivery in their present systems. Changes in program administration would be required
for all SDA's, though this was not a topic of study in this evaluation. The administrative
feasibility at the state level and its impact on other commodity donation program remain
major unanswered questions.
CLOC SFA manager views. Food service managers of CLOC SFA's reported
that the CLOC modifications had comparatively little adverse effect on their operations.
Increased administrative costs associated with dealing with changes in the use of full and
partial credit was the most frequent complaint followed by some dissatisfaction over the
purchase period for some commodities and over the discontinued use of refusals and
rollovers. Dissatisfaction with the purchase period appears to be associated with those
purchases that were required to be made very early in the year, before school started.
Domestic origin. An on-site survey of CLOC purchased inventories held by 19
CLOC SFA's in the spring of 1991 revealed that the labels on a significant portion of two
of the foods surveyed - canned tuna and apple juice -- indicated that they contained
vii
nn
imported product. This occurred despite SFA attempts to require distributors to provide
only domestically produced foods. It is concluded that: (a) the CLOC approach cannot
ensure that foods purchased are entirely of domestic origin and (b) that the incidence of
CLOC purchases of foods that are not of domestic origin is highest when USDA purchases
foods that are difficult to buy through commercial vendors or are significantly more
expensive than comparable foods of foreign origin.
Study Limitations
Sample Size and Representativeness. The size of the evaluation sample -- 25
CLOC systems and 25 Commodity systems - is small relative to the size of the universe of
over 15,000 SFA's. While this sample includes districts from various parts of the country
with varying enrollment sizes in both rural and urban settings, it was not designed to be
nationally representative. A limited comparison of sample district characteristics with
national totals indicates that the sample over-represents larger school districts and those that
participate in the School Breakfast Program. However, the CLOC sites and comparison
sites of Commodity SFA's are very closely matched to each.
Commodities Studied. The scope of the agricultural impact analyses was limited
to 13 major commodities. While these commodities accounted for approximately 75 percent
of the total value of entitlement commodities, restricting the evaluation in this way may have
overlooked potential market impacts for smaller purchase programs.
Administrative Feasibility. Questions on the administrative feasibility of the CLOC
system were limited to the ability of CLOC SFA's to adapt to the CLOC modifications and
to the effects of the modifications on local program administration. Issues related to
administration of the CLOC program at the State or National levels were not addressed.
viii
X
CONTENTS
Page
Executive Summary i
List of Tables xii
List of Figures xvii
Acknowledgements xviii
I. Introduction 1-1
A. History of the Commodity Donation Program 1-1
B. Program Operations 1-3
C. Results of Past Evaluation MO
D. Modifications of the Commodity Letter of Credit 1-13
II. Objectives of the Evaluation II-l
A. Issues to be Examined II-l
B. Research Questions II-2
III. Evaluation Methodology III-l
A. Evaluation Sample III-l
B. Data Collection III-2
C. Analytic Techniques III-8
D. Special Measurement Issues and Problems Ill-12
IV. Participating School Food Authorities IV-1
V. Evaluation Findings V-l
Questions:
1 Weight of commodity removed per KLEQ --
CLOC/Commodity compared V-l
2 Value of mix of products removed from the market --
CLOC/Commodity compared V-5
3 Share of value of foods attributable to raw
commodity value -- CLOC/Commodity compared V-19
4 Share of value of CLOC's net spent V-22
ix
X
Page
5 Share of entitlement received by Commodity SFA's .... V-28
6 CLOC issuance/commodity donation compared per
reimbursable meal V-38
7 Weight of commodity per dollar of CLOC issued «
1989/90 and 1990/91 compared V-41
8 Share of value of CLOC issuance not spent -
1989/90 and 1990/91 compared V-45
9 Share of CLOC's refused, rolled over, traded,
or not spent in 1989/90 compared to share traded
or not spent in 1990/91 V-46
10 Timing of CLOC and USDA purchases compared ..... V-52
11 Timing of CLOC deliveries and donated commodity
deliveries compared V-53
12 Pattern of CLOC purchases by month following
date of issuance -- 1989/90 and 1990/91 compared V-60
13 Changes in labor cost, labor hours, and food
costs per KLEQ -- CLOC/Commodity in 1989/90 and
1990/91 compared V-64
14 Rates of labor use by CLOC SFA's over time --
1983/84-1984/85 and 1990/91 compared V-66
15 Changes in participation rates ~ CLOC/Commodity in
1989/90 and 1990/91 compared V-71
16 CLOC manager's views of program modifications V-75
17 Effect on other programs of nationwide
adoption of CLOC V-77
18 USDA commodity purchasing criteria and practices .... V-86
19 USDA and commercial prices compared V-94
20 Domestic origin of CLOC purchases V-96
y
Page
21 Effect of CLOC modifications on program
administration V-103
VI. Study Limitations VI-1
A. Sample VI-1
B. Data VI-1
C. Methodology VI-3
VII. Conclusions VIM
Appendix A: Participating School Food Authority
Descriptive Data A-l
Appendix B: Comparison of Differences in Products
Acquired and Used by CLOC and Commodity
SFA's, SY 1990/91 B-l
xi
LIST OF TABLES
Page
I. Introduction
Table 1-1 Federal Support for School Food Programs,
FY 1969 - FY 1990 1-4
Table 1-2 Food Donated to School Districts Through The
Commodity Distribution Program, School Year 1990/91 1-6
II. Objectives of the Evaluation
Table II-1 Objectives and Research Questions II-2
m. Evaluation Methodology
Table III-l School Food Authorities Participating in the Commodity
Letter of Credit Modification Evaluation III-3
Table III-2 Correlation Coefficients of Variables Used in the
Regression Model III-ll
IV. Participating School Food Authorities
Table IV-1 Summary of Program Activity: 1989/90 IV-2
Table IV-2 Summary of Program Activity: 1990/91 IV-3
Table IV-3 Comparison of Participating SFA's on Selected
Variables By Demonstration Alternative
With National Totals IV-4
Table IV-4 Average Price Charged for Full Price Meal
and Change in Average Meal Prices,
1989/90 to 1990/91 IV-8
xu
Xlll
Page
V. Evaluation Findings
Table V-l Comparison of the Mean Farm-Equivalent Weight of
Commodities Acquired per KLEQ by CLOC and
Commodity SFA's, SY 1990/91 V-3
Table V-2 Comparison of the Mean Farm-Equivalent Weight of
Commodities Used per KLEQ by CLOC and
Commodity SFA's, SY 1990/91 V-4
Table V-3 Comparison of Differences in the Regression
Adjusted Mean Farm-Equivalent Weight of
Commodities Acquired and Used per KLEQ
by CLOC and Commodity SFA's SY 1990/91 V-6
Table V-4 Comparison of Differences in the Value of Products
Acquired as a Percent of the Total Value within
Each Commodity by CLOC and Commodity SFA's,
SY 1990/91 V-12
Table V-5 Comparison of Differences in the Mean Percent
Value of Acquired Products Derived From Each
Commodity Attributable to Raw Commodity Value
for CLOC and Commodity SFA's, SY 1990/91 V-20
Table V-6 Comparison of Differences in the Mean Percent
of Value of Used Products Derived From Each
Commodity Attributable to Raw Commodity Value
for CLOC and Commodity SFA's, SY 1990/91 V-21
Table V-7 Comparison of Differences in the Regression Adjusted
Mean Percent Value of Products Derived From Each
Commodity Attributable to Raw Commodity Value
for CLOC and Commodity SFA's, SY 1990/91 V-23
Table V-8 Mean Percent of the Total Value of CLOC Issuances
in SY 1990-91 Not Spent V-29
Table V-9 Comparison of the Distribution of USDA's
Reported Purchases, Deliveries to SDA's, CLOC
Issuances, and Donated Commodity Deliveries to
Study Sites, 13 Selected Commodities, SY 1990/91 .. V-33
Table V-10 Proportion of Total Entitlement Allocated to Selected
Commodities for CLOC SFA's, SY 1990/91 V-34
xiii
XW
Page
Table V-ll Mean Percent of the Value of Commodity
Entitlement Received by Commodity SFA's,
SY 1990-91 V-35
Table V-12 Comparison of the Mean Value of CLOC Issuances
per KRM and the Mean Value of Commodity
Donations per KRM, by Commodity, SY 1990/91 ... V-40
Table V-13 Comparison of the Mean Value of CLOC Issuances
per KRM and the Mean Value of Commodity
Donations per KRM, by Commodity, SY 1990/91 ... V-42
Table V-14 Comparison of the Farm Equivalent Weight of
Commodity Removed from the Market per Dollar of
CLOC Issuance in SY 1989-90 and SY 1990-91 V-44
Table V-15 Comparison of the Average Percent of CLOC Issuance
Not Spent in SY 1989/90 and in SY 1990/91 V-47
Table V-16 Comparison of the Mean Percent of the Value
of CLOC's Not Used in SY's 1989/90 and 1990/91 .. V-50
Table V-17 Comparison of the Date of CLOC Issuance and
Expenditure with the Date of USDA Purchase,
SY 1990-91 V-54
Table V-18 Mean Share of Donation Deliveries to SFA's
by Month, by Commodity, SY 1990/91 V-57
Table V-19 Comparison of the Date of CLOC Issuances
and Expenditure, SY 1989-90 V-61
Table V-20 Comparisons of Mean Changes in SFA Labor Hours,
Labor Costs and Food Costs Per KLEQ for CLOC and
Commodity SFA's, SY 1989/90 and SY 1990/91 V-65
Table V-21 Comparison of Differences in the Regression
Adjusted Mean Changes in SFA Labor Hours,
Labor Costs and Food Costs per KLEQ
Between SY 1989/90 and SY 1990/91 V-66
Table V-22 Comparison of the Number of LEQ Per Labor Hour
for CLOC SFA's in SY's 1983/84 - 1984/85
and SY 1990/91 V-70
xiv
vv
Page
Table V-23 Comparison of Rates of Participation and Changes
in the Rates of Participation In CLOC
and Commodity School Feeding Programs
Between SY 1989/90 and SY 1990/91 V-72
Table V-24 Comparison of Regression Adjusted Changes
in the Rate of Participation in CLOC and
Commodity SFA's between SY 1989/90
and SY 1990/91 V-73
Table V-25 Summary of CLOC SFA Managers's Opinions
About the Impact of CLOC Modifications V-75
Table V-26 CLOC SFA Manager Opinions Regarding the Effect
of CLOC Modifications on Operations,
by Commodity, SY 1990/91 V-77
Table V-27 Value of Domestic Food Donations, by Program,
1985/86-1990/91 V-80
Table V-28 Commodity Donation Program Responsibilities
of State Agencies Administering Commodity
Donations Through School Lunch Program V-81
Table V-29 Number of SDA's and Share of Shipment by Type of
Distribution System for NSLP SDA's,
1985/86 and 1989/90 V-83
Table V-30 Farm Gate Prices: 1989-90 and 1990-91
Marketing Years V-88
Table V-31 Volume and Value of USDA Commodity Donation
Deliveries through the NSLP, SY's 1989/90
and 1990/91 V-90
Table V-32 Comparison of USDA Purchases Prices for Selected
Products, 1990/91 V-93
Table V-33 Comparison of U.S. Domestic Production and Imports
of Selected Food Items, CY 1988 and CY 1989 V-96
Table V-34 Results of On-Site SFA Visits to Identify
Foreign Origin Foods V-99
xv
y\ji
Page
Table V-35 Allocation of Staff Time in Administering the
CLOC Program, SY 1989/90 and SY 1990/91 V-101
Appendix A. SFA Descriptive Data
Table A-l Size Distribution of SFA's Participating
in the Evaluation Compared to Size Distribution
Nationally in SY 1988/89 A-l
Table A-2 Food Service Program Responsibility
1989-90 School Year A-2
Table A-3 Food Service Program Responsibility
1990-91 School Year A-3
Table A-4 Number and Mean Share of Schools by Type
1989-90 School Year A-4
Table A-5 Number and Mean Share of Schools by Type
1990-91 School Year A-5
Table A-6 Participation Measures 1989-90 School Year A-6
Table A-7 Participation Measures 1990-91 School Year A-7
Table A-8 Reimbursable Lunches and Breakfasts
1989-90 School Year A-8
Table A-9 Reimbursable Lunches and Breakfasts
1990-91 School Year A-9
Table A-10 Food Service Options 1989-90 School Year A-10
Table A-11 Food Service Options 1990-91 School Year A-11
Table A-12 Kitchen Configurations 1989-90 School Year A-12
Table A-13 Kitchen Configurations 1990-91 School Year A-13
Table A-14 Average Meal Prices 1989-90 School Year A-14
Table A-15 Average Meal Prices 1990-91 School Year A-15
xvi
yvi
Page
Table A-16 Federal Lunch Reimbursement
1989-90 and 1990-91 School Year A-16
Table A-17 Staffing, School Year 1989-90 A-17
Table A-18 Staffing, School Year 1990-91 A-18
Table A-19 Sources of Income 1989-90 School Year A-19
Table A-20 Sources of Income 1990-91 School Year A-20
Table A-21 Food Service Expenses 1989/90 School Year A-21
Table A-22 Food Service Expenses 1990/91 School Year A-22
LIST OF FIGURES
Figure V-l Distribution of SFA's by Percent of
Entitlement Received V-37
Figure V-2 Overall Comparison of CLOC and Commodity
Delivery Patterns, SY 1990/91 V-58
Figure V-3 Apple Product Deliveries V-59
Figure V-4 Green Bean Product Deliveries V-59
xvu
y\/u\
ACKNOWLEDGEMENTS
A number of people contributed to the preparation of this report. Without attributing
to them any responsibility for errors or omissions, the authors would like to thank one and
all for their help.
A very special thanks is due the food service directors and staffs of the 50 school
systems that participated in the evaluation by supplying reams of information throughout
the year. Without their dedication to improvements in school food service, as evidenced by
their willingness to voluntarily take part in this demonstration, this evaluation wouldn't have
been possible.
Two other individuals who deserve special thanks for the contributions they made
throughout the evaluation are Cherie Root and Ann Hallawell of Virginia Polytechnic
Institute and State University. Having administered the CLOC pilot program for the past
10 years, the rich and unique experience they brought to the evaluation was of great value.
Beyond supervising the collection and transcription of the thousands of food acquisition and
inventory records used in the analysis, they served as general consultants throughout the
project.
The evaluation also benefitted from the periodic advice of Michael J. Puma of Abt
Associates, Dorothy V. Pannell of Fairfax County (Virginia) Public Schools and Joanne L.
Styer of Montgomery County (Maryland) Public Schools. Their advice, both practical and
methodological, based on their extensive experience in school feeding and in the evaluation
of school feeding programs was greatly appreciated.
As in most studies, this evaluation has been constructed on a foundation of knowledge
developed through earlier research. Two studies were particularly relevant and valuable in
this regard. The initial evaluation of CLOC by Abt Associates, completed in 1986, provided
a useful frame of reference, both for design of the evaluation and for comparison of results.
V Robert G. St. Pierre and Frederic B. Glantz, two of the principals in the original Abt
Associates evaluation, also made useful suggestions regarding the measurement of lunch
equivalents. A second report of particular value in preparation of the data analysis and
collection plan was prepared by Sheena McConnell and Harold Beebout of Mathematica
Policy Research. Inc.2/
" St. Pierre, Robert G. et. al., Evaluation of Alternatives to Commodity Donation in the
National School Lunch Program: Final Report. Abt Associates, Cambridge, Mass.,
prepared for the Food and Nutrition Service, USDA, 1986.
V McConnell, Sheena and Beebout, Harold, Analysis Plan for the CLOC Modification
Demonstration. Mathematica Policy Research, Inc.. Washington, D.C., prepared for the
Food and Nutrition Service, USDA April 3, 1990.
xviii
KIY.
We are indebted to those staff members of the Agriculture Marketing Service
responsible for commodity procurement who took time to discuss their procedures with us.
Our appreciation also to Alberta Frost, Director of the Food Distribution Division of FNS
for her advice and help and to the FNS Regional Office Staff who collected information for
analysis of the domestic content issue.
Sheila Wilson of Abel, Daft & Earley maintained her cheerful composure throughout
the typing of several drafts of this report, for which we express our appreciation.
Last but certainly not least, John R. Endahl, the FNS Contracting Officer's
Representative for the project, has our special thanks. Having participated in both the
administration and the evaluation of the CLOC demonstration for several years, he brought
to the project a balance of experience, insight, and analytic competence that was of
substantial value. He provided critical advice and support throughout the project.
xix
yx
L INTRODUCTION
This report presents the findings of an extensive evaluation of modifications in the
Commodity Letter of Credit (CLOC) program that were adopted by school districts taking
part in the program in School Year (SY) 1990/91. For the past ten years, beginning in SY
1982/83, 25 school districts throughout the United States have participated in a
demonstration ofCLOC as an alternative to the U.S. Department of Agriculture's (USDA's)
Food Distribution Program to schools in the National School Lunch Program (NSLP).
Under the program, school districts are authorized through letters of credit to make their
own purchases of specified foods in place of receiving donated commodities purchased by
USDA. Results of an earlier evaluation, ending in 1986, led the USDA to conclude that the
CLOC approach did not warrant continuation. However, at the urging of school districts
taking part in the demonstration and other associated interests, the Congress authorized the
participating districts to continue in the CLOC program. On three other occasions since
then the program has been extended and a measure is now pending in Congress to extend
it once again. The issue of whether CLOC should be adopted on a broader scale or
discontinued altogether remains unresolved.
In the Conference Report language accompanying the Child Nutrition and WIC
Amendments Act of 1989, the Congress authorized modifications in the CLOC program as
a means of determining whether the most commonly voiced criticisms of the program could
be overcome while, at the same time, retaining CLOC's principal advantages. The central
purpose of this evaluation is to examine the extent to which these modifications have
achieved that purpose.
The report begins with a brief description of the background that led to the design
of CLOC, its implementation on a demonstration basis, and results of the initial evaluation.
Shortcomings of the program are discussed, as are the modifications that were adopted in
1990/91 to address these limitations. This is followed by a description of the methodology
employed in conducting the study and a profile of the school systems that participated in
it. Study findings are then reported for each of the research questions that are being
examined. A final chapter presents major conclusions of the evaluation.
A. History of the Commodity
Donation Program
Beginning in the 1930's when farm prices were severely depressed and the ranks of
the unemployed were growing ever larger, the Federal government began donating surplus
food to those who were poor and unemployed. Donation programs were viewed then, as
they continue to be viewed today, as an opportunity to improve the nutritional status of
people in need while at the same time strengthening the demand for commodities that are
temporarily in surplus, and in that way supporting farm prices and farm income.
Since their first use nearly 60 years ago, donated commodities have been used to
achieve several objectives and to reach different populations. For example, the commodity
donation programs that began in the 1930's were the antecedents of the current Food Stamp
1-1
Program, now the nation's largest food assistance program funded at over $17 billion in FY
1991 and serving over 22.6 million people. Schools first began receiving food donations in
the 1930's. However, it was not until the adoption of the National School Lunch Act of 1946
that the role of food assistance in the schools was formalized and provided on a regular
basis. Currently, the NSLP is by far the largest commodity donation program. However,
there are another 12 programs that serve such widely divergent populations as charitable
institutions, summer camps, and Indian reservations.
Debate over whether to provide food assistance in the form of increased purchasing
power (e.g., cash, food stamps, or letters of credit) or in the form of commodities has been
a part of the policymaking process for these programs since their earliest days. The initial
move tov/ard providing eligible food aid recipients with redeemable coupons instead of
donated foods began in May, 1939 when the U.S. Department of Agriculture's first
experimental food stamp program was initiated in Rochester, New York. & Still, it was not
until many years later, in August 1964 that a full-scale, nationwide Food Stamp Program was
established.
In the case of school food service programs, a combination of cash assistance and
commodity assistance has been provided since the beginning of the NSLP. During the early
years of the program, commodity donations played a more important role than they do now.
In FY 1965, commodities accounted for over two-thirds (68 percent) of Federal support.^
Over time, the relative importance of commodity assistance to schools has declined,
although it continues to be important to industry in surplus removal and market
stabilization. Still, donated commodities remain a significant source of support to school
food service programs throughout the nation. In SY 1990/91 schools participating in the
NSLP received donated foods valued at $624 million or about 17 percent of total Federal
assistance to the program.
Donated commodities are classified as "entitlement" or "bonus". Entitlement
commodities are distributed at a legislatively determined rate based on the number of
lunches served. In SY 1990/91, school districts were entitled to receive commodities valued
at 14 cents for every lunch served that met USDA requirements. Since SY 1981/82, the
entitlement rate has increased at an average annual rate of 2.8 percent. Bonus commodities
are made available as condition of crop surpluses warrant. Bonus commodities were first
distributed in FY 1978. With large surpluses of dairy products in the mid-1980's, bonus
commodities grew rapidly reaching $440 million in FY 1987 and accounting for 49 percent
of the value of all donated commodities through the NSLP that year. More recently bonus
commodities have been distributed in far smaller amounts. In SY 1990/91, $83 million of
bonus commodities were distributed.
U U.S. Congress, Senate Committee on Agriculture, Nutrition, and Forestry, The Food
Stamp Program: History. Description. Issues, and Options. April 1985, p.4.
^ Congressional Budget Office, Feeding Children: Federal Child Nutrition Policies in the
1980's May 1980, p.50.
1-2
Trends in the level of federal assistance to school feeding programs over the period
1969-90 are shown in Table 1-1.
As the nation's school systems consolidate and grow larger, as the technology and
resource requirements of meal preparation evolve, and as the nation's eating habits change,
there are new and different requirements for the types, forms, and unit sizes of foods that
schools can most effectively use as well as changes in the nature of the distribution system
through which these foods are made available to the schools. Within the farm sector,
however, the need for price stabilizing actions in commodity markets has remained, though
with the exception of the early 1980's, the farm economy has been far healthier than it was
during the Great Depression. For mo |