Food and Agricultural Commodity Consumption in the United States:…

Food and Agricultural Commodity Consumption in the United States:
Looking Ahead to 2020. By Biing-Hwan Lin, Jayachandran N. Variyam,
Jane Allshouse, and John Cromartie, Food and Rural Economics Division,
Economic Research Service, U.S. Department of Agriculture, Agricultural
Economic Report No. 820.




                                  Abstract
U.S. consumption of food commodities is projected to rise through the year 2020,
mainly due to an increase in population. But the mix of commodities is expected to
shift because of an older and more diverse population, rising income, higher
educational attainment, improved diet and health knowledge, and growing popu-
larity of eating out. This study analyzes data from USDA's food consumption
survey to project the consumption, through the year 2020, of 25 food groups and
22 commodity groups. Per capita consumption of fish, poultry, eggs, yogurt, fruits,
nuts and seeds, lettuce, tomatoes, some other vegetables, grains, and vegetable oils
is predicted to rise, whereas consumption of beef, pork, other meat, milk, cheese,
potatoes, and sugar is expected to fall. The growth of the at-home and away-from-
home markets varies from one commodity to another. Fruit consumption is
expected to lead all commodities in growth in the at-home market, and fish
consumption is expected to lead in growth in the away-from-home market.

Keywords: Eating out, diet and health knowledge, food-commodity translation
database, food consumption projections, commodity consumption projections, and
Continuing Survey of Food Intakes by Individuals, 1994-96 and 1998.




                         Acknowledgments
The authors thank Chung L. Huang of the University of Georgia, W. Noel Blisard,
Nicole Ballenger, and James Blaylock of the Economic Research Service for
helpful review of the earlier drafts of this report, Lou King for editorial advice,
Wynnice Pointer-Napper for the final document layout and charts, and Curtia
Taylor for cover design.




1800 M Street, NW
Washington, DC 20036-5831                                         February 2003
                                                                        Contents
                     Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
                     Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
                     Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
                     Model Specification, Estimation, and Forecasting . . . . . . . . . . . . . . . . . . . . .                                4
                      Indirect and Direct Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .                     4
                      Model Structure and Specification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .                      4
                      Estimation Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .               6
                     Estimation Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .          8
                        Eating-Out Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .              8
                        Diet-Health Knowledge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .                8
                        Food Consumption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .             9
                     Projections of Food and Commodity Consumption, 2000-2020 . . . . . . . . . . 11
                        Economic, Social, and Demographic Variables, 2000-2020 . . . . . . . . . . . . . 11
                        Assumptions Underlying the Forecasts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
                        Projections of Food Consumption, 2000-2020 . . . . . . . . . . . . . . . . . . . . . . . 14
                        Projections of Commodity Consumption, 2000-2020 . . . . . . . . . . . . . . . . . . 15
                     Decomposition and Sensitivity Analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
                       Effects of Economic, Social, and Demographic Factors . . . . . . . . . . . . . . . . 23
                       Effects of Eating Out and Diet-Health Knowledge on
                       Commodity Consumption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
                       Direct and Indirect Effects of Income . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
                       Sensitivity Analysis of Eating Out and Age . . . . . . . . . . . . . . . . . . . . . . . . . 26
                     Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
                     References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
                     Appendix tables: Tobit Results of Food Consumption Equations . . . . . . . . 34




ii      Food and Agricultural Commodity Consumption in the United States / AER-820                                             Economic Research Service, USDA
                                                    Summary
                 As the American population becomes older and more racially and ethnically
                 diverse, the volumes and types of foods preferred can be expected to shift. Older
                 folks, for example, tend to eat out less often than younger folks, and, to some
                 extent, prefer different foods. An increasing proportion of elderly in the popula-
                 tion, therefore, will have implications for the types of foods demanded. Likewise,
                 different ethnic subpopulations have some distinct food preferences.

                 This report examines the volumes of individual types of foods eaten by Americans in
                 the years 1994-98, and projects those volumes to the year 2020, taking into account
                 population and demographic change as well as trends in economics and immigration.

                 A two-step econometric system was specified to model food consumption. The
                 first step was to estimate consumers' eating-out habits and their diet-health knowl-
                 edge. In the second step, food consumption, both at home and away from home,
                 was related separately to consumers' eating-out habits, knowledge, income, and
                 their social and demographic characteristics. The Tobit procedure was utilized to
                 deal with the fact that many consumers may not consume certain foods during a
                 survey period. Using projected values of economic, social, and demographic
                 factors for 2000-2020, we projected at-home and away-from-home food consump-
                 tion for the same period. We then developed a food-commodity translation data-
                 base to convert food consumption to commodity consumption. Twenty-five food
                 groups and 22 commodity groups were analyzed in this study.

                 This study makes several contributions to the literature on food consumption. To
                 our knowledge, no research has been conducted to model the effects of both eating
                 out and diet-health knowledge on food consumption. Likewise, the at-home and
                 away-from-home separation of food consumption had never been attempted in the
                 literature. We developed a food-commodity translation database to convert food
                 consumption to commodity consumption. This translation database contains the
                 amounts of several hundreds of commodities in each of more than 7,000 food items.

                 Due to an anticipated population growth of 50 million between 2000 and 2020 in
                 the United States, total consumption of all 22 commodities is predicted to rise,
                 even though per capita consumption of many commodities is predicted to fall. The
                 results suggest that fruits will lead all commodities in terms of growth in both total
                 and per capita consumption. Certain vegetables, such as lettuce and tomatoes, are
                 predicted to grow substantially, while per capita potato consumption (fried and
                 other) is predicted to decline, slowing down the growth in total U.S. potato
                 consumption. Increases in meat, poultry, and fish consumption vary. Per capita fish
                 and poultry consumption is predicted to rise while beef, pork, and other meat per
                 capita consumption is predicted to fall. Per capita consumption of milk and cheese
                 is predicted to fall, while per capita consumption of yogurt and eggs is predicted to
                 rise. The consumption of nuts and seeds and grains is also predicted to rise over
                 the next two decades.

                 The consumption projections are based on expected shifts in economic, social, and
                 demographic conditions. In addition, the consumption projections are sensitive to
                 two underlying assumptions. First, no prices or expenditures on food consumption
                 were reported in the surveys, so relative prices are assumed to remain constant
                 during the survey period (1994-96 and 1998) and the projection period (2000-


Economic Research Service, USDA                Food and Agricultural Commodity Consumption in the United States / AER-820      iii
                    2020). Second, there is an implicit assumption that as an individual moves from
                    one demographic group to another, his/her preferences immediately take on the
                    characteristics of the new group. A sensitivity analysis was conducted to examine
                    the effects of assuming different eating-out habits as adults grow older between
                    2000 and 2020. The results suggest that the projections of some commodities, such
                    as potatoes, are quite sensitive to the assumption about eating-out habits.




iv      Food and Agricultural Commodity Consumption in the United States / AER-820      Economic Research Service, USDA