Commercial Market Tracking Systems: Applications to Policy Formation Regarding Nutrition Labeling

Michael L. Stewart, Food and Drug Administration
ABSTRACT - The first part of this paper describes the construction of a data base integrating retail market share and food package information for a sample of packaged processed foods. The data base was then employed to estimate the extent of nutrition labeling in the retail food market as well as the amount of nutrition labeling carried out voluntarily and the degree to which food manufacturers were violating the food labeling regulation by not including s nutrition label where required.
[ to cite ]:
Michael L. Stewart (1981) ,"Commercial Market Tracking Systems: Applications to Policy Formation Regarding Nutrition Labeling", in NA - Advances in Consumer Research Volume 08, eds. Kent B. Monroe, Ann Abor, MI : Association for Consumer Research, Pages: 488-492.

Advances in Consumer Research Volume 8, 1981      Pages 488-492

COMMERCIAL MARKET TRACKING SYSTEMS: APPLICATIONS TO POLICY FORMATION REGARDING NUTRITION LABELING

Michael L. Stewart, Food and Drug Administration

ABSTRACT -

The first part of this paper describes the construction of a data base integrating retail market share and food package information for a sample of packaged processed foods. The data base was then employed to estimate the extent of nutrition labeling in the retail food market as well as the amount of nutrition labeling carried out voluntarily and the degree to which food manufacturers were violating the food labeling regulation by not including s nutrition label where required.

BACKGROUND

Through the early and mid-1960's, emphasis in food marketing and purchasing was primarily limited to product taste, appearance, price, and convenience. Nutrition did not constitute a dominant (or even important) theme beyond the basic rubric of attempting to select at least one item from each of several basic food groups to obtain "a well-rounded meal". Beginning in the latter part of the 1960's, however, nutrition begin to acquire a more dominant position for both policy-makers and consumers. In 1967, the Senate Subcommittee on Employment, Manpower and Poverty uncovered widespread malnutrition and hunger in its travels through the Delta area of Mississippi. In 1967, the Citizen's Beard of Inquiry into Hunger and Malnutrition held that as of that time ten million people across the nation were suffering from hunger and malnutrition. These conclusions were further supported by the findings of the federal government's 1968 Ten-State Nutrition Survey. Therefore, in 1968 Congress established the Senate Select Committee on Nutrition and Human Needs to act as watchdog and advocate for government food program. In 1969, President Nixon called for a White House Conference on Food, Nutrition, and Health, and one panel was detailed to deal with food packaging and labeling (Austin 1977). This panel stipulated that the principal function of a food package is to protect the product and to provide an efficient means for its distribution, while the principal function of a food label is to identify and describe the product and to provide information to promote use of the product. The panel concluded that new packaging and labeling concepts for better nutrition and nutrition information would succeed to the extent that they supported these basic functions. Three of their recommendations were:

1. initiate a comprehensive review of laws and regulations to determine whether they insure and facilitate the delivery of sound nutrition to the consumer;

2. examine attitudes, economic barriers, ethnic preferences, geographic patterns, and other factors relevant to improving nutrition and health, including full utilization of the package and the label; and

3. appoint special committee to consider specific aspects of nutrition (e.g., ways to simplify labels while at the same time making them more descriptive nutritionally).

As a direct result of the White House Conference, the Food and Drug Administration (FDA) began to explore different methods of displaying nutrition label information. In 1972, the FDA presented a proposal for a nutrition labeling regulation which was modified and issued as a final regulation in 1973 (with an effective compliance date of July, 1975). The regulation emphasized voluntary rather than mandatory labeling in an attempt to speed acceptance and to avoid litigation which would delay implementation. The main elements of the regulation stated (Austin 1977, pg. 29):

1. the presence of a nutrition label is voluntary except when a nutrient is added to a processed food and/or a nutrition, claim is made;

2. the nutritional information, if present, must be presented conspicuously on or to the immediate right of the principal display panel;

3. the information presented must include serving size, servings per container, calorie/protein/carbohydrate/fat content, and the percentages of the U.S. Recommended Daily Allowances (U.S. RDA)for eight specified nutrients;

4. a claim cannot be made that a food is a significant source of a nutrient unless each serving provides at least 10% of the U.S. RDA of that nutrient, nor can a product claim nutritional superiority in a nutrient unless it contains at least 20% of the U.S. RDA for that nutrient;

5. imitation foods are required to be labeled as imitation only when they are nutritionally inferior;

6. cholesterol and fatty acid labeling is permitted but must be accompanied by the other nutrition label information;

7. samples of fabricated or fortified foods to which nutrients have been added must contain at least 100% of their labeled value, while samples of food in which there are naturally occurring nutrients must contain et least 80% of their labeled value; and

8. fortified foods containing from 50% to 150% of the U.S. FDA of any nutrient are considered to be a dietary supplement rather than an ordinary food and are thus subject to additional regulations, while food with U.S. FDA content over 150% falls into the drug classification.

Food industry response to the labeling regulation ranged from confrontation to formulation of nutrition policies and organization of new consumer departments. Throughout the 1970's the pace of nutrition-oriented product development accelerated, with the appearance of "natural" foods and soy-based textured vegetable products. Various members of the food industry utilized the new consumer awareness of nutrition as a marketing strategy to expand their market shares. Other members of the food industry felt obliged to use nutrition labeling as a defensive marketing strategy to prevent the loss of old customers who would otherwise switch to nutrition-labeled brands (Austin 1977).

Data-based estimates of the extent of nutrition labeling are limited. The Grocery Manufacturers of America (GMA 1975) conducted a survey of 110 member and 155 non-member companies prior to the 1975 effective date for the nutrition libeling regulation. Of the 79 companies that responded, most were food manufacturers, with a few retailers included. Eighty-five percent of those manufacturers responding indicated an intent to label some if not all of their products, Of the 15% responding negatively, 58% felt that their products were inappropriate for labeling (e.g., apple butter, fresh meat, margarine, cake flour, jellies, mustard, etc.). In nearly every product category, some manufacturers planned to institute nutrition labeling, and some did not. Larger companies tended to plan to institute nutrition labeling sooner than did smaller companies.

Austin (1977), for the Market Science Institute, integrated the GMA survey and other sources to conclude that $17.5 billion (approximately 14%) of the $127 billion representing total U.S. food consumption (including away-from-home consumption) was nutritionally labeled. Furthermore, approximately 30% of the $57.6 billion remaining after exclusion of away-from-home consumption, condiments, alcoholic beverages, and fresh fruits and vegetables was nutritionally labeled.

Because of the scarcity of direct data estimating the total incidence of nutrition labeling and because of the importance of overseeing and perhaps modifying the nutrition labeling regulations, the Division of Consumer Studies, Bureau of Foods, FDA implemented the formation of the "Food label and Package Surveillance (FLAPS) Data Base". Five objectives were formulated (Schucker 1978):

1. develop a sales-based estimate of the volume-importance of packaged foods currently sold under nutrition labeling;

2. extend the nutrition labeling conclusions to individual product classes;

3. estimate the percentage of packaged foods voluntarily labeled and assess the food industry initiatives in labeling;

4. determine the quantity and volume-importance of retail food sold in violation of nutrition labeling regulations;

5. make the data base open-ended to allow examinations of trends across years.

Data collection began in 1977.

METHODOLOGY

The FLAPS data base integrates two data sources: (1) the A. C. Nielsen Directory of Supermarket Products and (2) FDA in-house compilations of package and package-label information, off-label advertising, and package construction materials. The Nielsen Directory provided the sampling frame and market information. The in-house compilation supplied all non-market information (label information, package construction, package dimensions, etc.).

Nielsen Directory - Sample Specification

The sampling universe for the initial round of the FLAPS data base (FLAPS-1) consisted of the A. C. Nielsen Directory of Supermarket Products for April-May, 1975-76. The Nielsen Directory, part of the Nielsen Early Intelligence System (NEIS), is a syndicated sales data base concerned with merchandise movement from central warehouses to a selected sample of U.S. food stores. The information in each directory represents bimonthly statistics for two consecutive years. Weights furnished by A. C. Nielsen enable the user to project the bimonthly sample to estimate total annual retail movement in the United States. The weights are based upon: (l) the ratio of number of stores in the sample to total Nielsen-type stores in the U.S.; (2) the percent of foods in a particular product class that move through Nielsen-type stores; and (3) the fraction of the total year during which the sample was collected. The information listed in the Nielsen Directory for any single item line within a brand includes can size, average retail price, and unit and dollar movement. The market share information is aggregated to the brand and product class levels. The 411 product classes included in the Nielsen Directory were aggregated by Nielsen at the request of the FDA into 51 major supermarket food groups (e.g., product classes "Spaghetti - Dry" and "Macaroni - Dry" became members of the product group "Pasta - Dry"). Sampling consisted of a multi-stage selection process.

Product class Selection.  Product classes within each product group were rank-ordered by decreasing retail dollar sales volume. Those product classes within a product group with largest dollar volume were selected into the data base until two-thirds of the retail dollar volume for each (and every) product group was accounted for. This procedure resulted in the selection of 158 of the 411 product classes. Forty additional product classes were randomly selected from the remaining 253, resulting in a total of 198 product classes, accounting for 852 of the total retail dollar volume of the 51 product groups.

Brand Selection.  Six national brands were drawn from each of the 198 product classes: (1) those three brands in each product class ranked first through third (market leaders) in terms of retail dollar sales and (2) three additional brands selected randomly from the remaining brands (non-leaders) in each product class. (If six or fewer brands were available for any one product class, all brands were selected.)

Due to restrictive agreements between the A. C. Nielsen Company and retail organizations, sales information for individual private label brands was not included in the Nielsen Directory. Therefore, a modified sampling plan for private label brands was instituted, such that those 20 product classes with the largest private label retail dollar sales volume were selected. Twenty additional product classes were randomly selected from those remaining product classes that contained private label brands. Based upon a separate listing of retail firms with one or more private label brands, those three firms with the largest total annual retail food sales and three randomly selected firms were chosen in each product class. Field personnel were permitted to pick up any private label brand in a specific product class controlled by the sampled firm.

Item Selection.  A "brand" was defined to include all sizes, flavors and form variations of a product under a particular label, selection of a specific item within a brand (e.g., Brand X whole tomatoes in can size 303 rather than Brand X peeled tomatoes in can size 300) was accomplished by selecting the item with the largest sales volume within that brand. It was decided that those items would be most likely to reflect the manufacturers' current labeling policies and most recent label formats.

Special Samples - Meat and Convenience Foods.  Processed meat and poultry and products containing meat and poultry, although regulated by the U.S. Department of Agriculture (USDA), were included in the FLAPS data base to determine the current extent of nutrition labeling for those products in anticipation of the issuance of labeling regulations by the USDA similar to those of the FDA. To segregate the products from those regulated by FDA, a separate product group was created. The sampling procedure was identical to that for FDA-regulated products.

Convenience samples (store-door delivery) of bread, fresh milk, and ice cream products were collected in three major markets: New York, Chicago, and Los Angeles metropolitan areas. All brands encountered in two leading chain stores and one leading independent store in each market were purchased. This sampling procedure was instituted due to lack of available retail sales information at the brand level for this category.

In-House Data Collection

Field personnel collected 1,713 individual brand items between April, 1977 and March, 1978. These items were sent to the Washington, DC office of the FDA for processing. Products without nutrition labels were sent as empty packages. Products with nutrition labels were sent intact (unopened) and were assayed for nutrient content upon arrival. All information on package labels (including all nutrition label information) was recorded. Access was obtained to the food products advertising file of the Bureau of Consumer Protection of the Federal Trade Commission. All available magazine and newspaper advertisements, radio scripts, and television commercials (in storyboard form) were analyzed for content and classified as to presence or absence of health and/or nutrition claims that would trigger mandatory nutrition labeling. When all information to be compiled in-house had been prepared, the resulting file was merged with the Nielsen marketing data and set up as a computer file data base.

FLAPS-1 Sample Coverage

The 51 product groups (all product classes) in the 1975-76 Nielsen Directory represented an estimated $45.9 billion in annual retail food sales. Excluding meat and meat containing products (regulated by the USDA), milk, milk products, snacks, and carbonated soft drinks (because of the significant amount of store-door delivery), remaining food products accounted for $38.1 billion (360 product classes).

The FLAPS-1 data base (excluding meat and meat containing products, milk, milk products, snacks and carbonated soft drinks) totaled $22.3 billion in projected annual retail sales (158 product classes, or 1,055 brands). Employing non-leader brands as unbiased estimators for all non-leaders in sampled product classes, the $22.3 billion was projected to $27.3 billion. The 862 market leader and non-leader brands, all of which had brand level sales data available, represented $23.1 billion of the $27.3 billion in retail market sales.

RESULTS

The data in FLAPS-1 were examined using two dependent measures: (1) percent of brands (to estimate the extent' of nutrition labeling among manufacturers) and (2) share of market dollars (to estimate the impact of labeling practices upon the retail market). Unless specifically noted, all analyses and figures are based upon packaged processed foods, excluding processed meat and meat-containing products, milk, milk products, snacks and carbonated soft drinks.

General Analysis

Approximately one third (33.5%) of the 862 national brands in the FLAPS-I data base displayed a nutrition label (see Table 1). These nutrition labeled brands

TABLE 1

PERCENT OF PROJECTED ANNUAL RETAIL MARKET DOLLARS (AND PERCENT OF BRANDS) FOR NATIONAL BRANDS

accounted for 47.5% of the $23.1 billion dollars in projected annual retail sales for national brands sampled by FLAPS-1. Approximately the same proportion of brands displayed nutrition labels (16.1%) solely on a voluntary basis as fell under the mandatory nutrition labeling requirement (17.4%) because of either an on-label nutrition claim, an off-label nutrition claim, the addition of one or more nutrients to the product, or some combination of the above. However, those brands with required nutrition labels accounted for a greater relative share of total market dollars (27.6%) than did brands with voluntary nutrition labels (19.9%) and share of market for both was greater than the corresponding percent of brands falling into those two categories. Brands in violation of nutrition labeling regulations (required label not present) represented 0.9% of all brands and 0.6% of all retail market dollars. Over half of all brands (66.5%) were not required to and chose not to display a nutrition label (but accounted for only 52.5% of the retail market).

Market leader vs. Non-leader Brands

Labeled vs. unlabeled.  Approximately equal numbers of market leaders (N = 460) and non-leaders (N = 402) were sampled in FLAPS-1 (see Table 2). Market leaders accounted for a greater relative share of market dollars ($17.2 billion) than non-leaders ($5.9 billion), as would be expected given the definition of leaders and non-leaders: leaders represented the three largest brands in each product class in terms of retail dollar sales. A greater proportion of leader brands (40.7% of the total number of leader brands) than non-leader brands (25.4% of the total number of non-leader brands) engaged in nutrition labeling. For both leaders and non-leaders, the share of total market dollars accounted for by nutrition labeling (51.6% and 35.7%, respectively) was greater than the corresponding proportion of all leader and non-leader brands with nutrition labeling (40.7% and 25.4%, respectively).

Mandatory vs. voluntary labeling.  Market leader brands with nutrition labels were divided evenly between brands with required nutrition labels (20.0%) and those engaged in voluntary nutrition labeling (20.7%). Market leaders with required nutrition labels accounted for a larger proportion of total market leader dollars (29.3%) than did market leaders engaged in voluntary nutrition labeling (22.3%). In both categories, however, share of marker was greater than the relative proportion of brands engaged in nutrition labeling.

More non-leader brands engaged in mandatory (14.4%) than in voluntary nutrition labeling (10.9%). Furthermore, the share of market dollars accounted for by mandatory labeling among non-leaders (22.8%) was more than a proportionate increase over the share accounted for by voluntary labeling among non-leaders (12.8%).

TABLE 2

PERCENT OF PROJECTED RETAIL MARKET DOLLARS (AND PERCENT OF BRANDS) FOR MARKET LEADERS, NON-LEADERS, AND PRIVATE LABELS

Labeling regulation violations.  Approximately the same proportion of leader (0.9%) and non-leader (1.0%) brands did not display a required nutrition label. In both instances, share of market dollars (0.7% and 0.4%, respectively) was less than the percent of brands.

Private Label Brands

Nutrition labels were present on 43.5% of all private label brands (a greater overall proportion than for either market leaders or non-leaders). In contrast to market leaders and non-leaders, however, the proportion of private label brands with required labels (19.2%) was less than the proportion with voluntary labels (24.4%). Because of the lack of private label brand-level sales information, share of market dollars for private label brands (see Table 2) was not comparable to share of market dollars for leaders and non-leaders. Total private label dollars within a product class had been divided equally among all private label brands within that product class for purposes of analysis and thus did not reflect individual brand share differences.

Fewer private label brands violated the nutrition labeling regulations (0.5%) than either market leader (0.9%) or non-leader brands (1.0%). At the same time, more private label brands engaged in voluntary labeling (24.4%) than either market leader (20.7%) or non-leader brands (10.9%).

Association Between Labeling and Share of Market

The fact that nutrition labeled brands (in aggregate) accounted for a greater relative proportion of the retail market dollars than did non-labeled brands (and mandatory more than voluntary) was indicative of a possible association between presence (and type) of labeling and brand share of market. A point-biserial correlation for all market leader and non-leader brands produced a modest but significant positive correlation between presence/absence of a nutrition label and brand share of market (r = .14, p<.01, n = 862). A point-biserial correlation between mandatory/voluntary labeling and brand share of market again produced a modest but significant correlation (r = .11, p < .01, n = 862). However, the mandatory/voluntary labeling by brand share correlation was not significant when limited to only those brands displaying a nutrition label (r = .05, p > .10, n = 289).

Mandatory Nutrition Labeling Triggers (National Plus Private Label Brands)

Labels present.  Of the 1,055 national and private label brands in FLAPS-1, 196 were required by regulation to display a nutrition label. The most frequent single trigger for labeling (N = 86) for those brands that did display a nutrition label (N = 187) was an off-label nutrition claim (see Table 3). The two remaining triggers, added nutrients (N = 35) and an on-label nutrition claim (N = 30) occurred with about equal frequency. Nutrition labels on the remaining 36 brands were triggered by combinations: an on-label nutrition claim plus added nutrients (N = 7) and an on-label nutrition claim plus an off-label nutrition claim plus added nutrients (N = 29).

Labels absent.  For the nine labeling violations, the most frequent causes were on-label nutrition claims (N = 4) and on-label nutrition claims plus added nutrients (N = 3). The last two violations were caused by added nutrients (N = 1) and an on-label nutrition claim plus an off-label nutrition claim plus added nutrients (N = 1).

USDA Regulated Products

A cross-tabulation was performed for only those 210 products in FLAPS-1 regulated by USDA (see Table 4). If the products had been subsumed under FDA regulations, 2.9% of the 210 brands would have been required to display package nutrition labels. Approximately one-third of those brands (1.0% of all 210 brands) did in fact display a nutrition label. The remaining two-thirds of the brands that would have been required to display a nutrition label (1.9% of the 210 brands) would have been in violation of the labeling regulation. Out of the 9.0% of the brands that did display nutrition labels (either required or voluntary), 8.1% would have been considered engaged in voluntary labeling. Finally, 89.0% of the brands did not display nutrition libels but would not thereby have violated the regulation.

TABLE 3

DISTRIBUTION OF FREQUENCY OF OCCURRENCE FOR TRIGGERS OF REQUIRED NUTRITION LABELING

TABLE 4

PROCESSED MEAT AND MEAT-CONTAINING PRODUCTS ONLY

CONCLUSION

An estimated one-third of all national brand processed foods (excluding meat, milk, milk products, snacks, and carbonated soft drinks) packaged in the United States in 1976 displayed a package nutrition label. These nutrition-labeled brands accounted for slightly less than one-half of the estimated total retail dollar volume for national brand products. This figure appears consistent with Austin's (1977) estimate of 30% for national retail sales under nutrition libeling, given the differences in sample composition. (The FLAPS-1 estimate excluded all private label brands, meat and meat containing products, milk, milk products, snacks, and carbonated soft drinks in addition to the away-from-home dining, condiments, alcoholic beverages, and fresh fruits and vegetables excluded by Austin.) Half of those brands with nutrition labels were required to display the information. Thus, one-half of nutrition labeled national brands (or one-sixth of all national brands) either made on-label or off-label nutrition claims, added one (or more) nutrients to the product, or engaged in a combination of all three. A further one-sixth of all national brands engaged in nutrition labeling on a purely voluntary basis (i.e., although the package contained a nutrition label, no on- or off-label nutrition claims were made and no nutrients were added to the product). Brands engaged in voluntary nutrition labeling accounted for approximately one-fifth of the estimated retail dollar sales for national brands in 1976. Brands with mandatory nutrition labeling, while approximately equal in number to brands with voluntary labeling, accounted for half again (roughly three-tenths) of estimated national brand retail dollars. Thus, share of retail dollar sales wars proportionately greater for nutrition-labeled than non-labeled brands and, within nutrition-labeled brands, for brands with mandatory rather than voluntary nutrition labeling. Significant albeit small positive correlations between (1) presence of nutrition labeling and share of market dollars and (2) type of nutrition labeling and share of market dollars supported the conclusion of a direct association.

A total of 0.9% of all national brands and 0.5% of all private label brands violated the nutrition labeling regulation by not displaying a required nutrition label. For national brands, the ratio of "share of retail dollars" to "percent of brands" in violation was less than 1.0. (Market share information at the brand level was not available for private label brands.) Thus, those manufacturers that violated the nutrition labeling regulation tended to be the smaller manufacturers. Label nutrition claims (with and without added nutrients) were the most frequent causes of the violations.

Whether nutrition labeling was instrumental in increasing brand share of market or whether large companies instituted nutrition labeling prior to smaller companies (or both) cannot be determined from the present data. However, the direct relationship between brand share and presence of labeling does appear consistent with the earlier GMA (1975) finding that large manufacturers intended to begin nutrition labeling before smaller companies. Later extensions of the FLAPS data base may provide insights as to the causal relationship between labeling and size of market share as trend data across time for both nutrition labeling and size of manufacturer become available.

REFERENCES

Austin, James E. (1977), "Food Marketing: What Role for Nutrition?", Marketing Science Institute, Report No. 77-104.

Grocery Manufacturers of America (1975), "Results of the GMA Survey on Nutrition".

Schucker, Raymond E. (1978), "A Surveillance of Nutrition Labeling in the Retail Packaged Food Supply," Division of Conair Studies, Bureau of Foods, Food and Drug Administration, Washington, DC.

White House Conference on Food, Nutrition and Health: Final Report, (1970), Washington: U.S. Government Printing Office.

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