A supply and demand estimation of the United States high...
Transcript of A supply and demand estimation of the United States high...
A supply and demand estimation of the United States high fructose corn syrup market
P. Lynn Kennedy, PhD
Crescent City Tigers Alumni Professor
Agricultural Economics and Agribusiness
Louisiana State University and LSU AgCenter
181 Martin D. Woodin Hall, Baton Rouge, LA 70803
Pablo A. Garcia-Fuentes, Ph.D.
Assistant Professor of Economics
Department of Economics, Finance, and General Business
Dillard College of Business Administration
Midwestern State University
3410 Taft Blvd
Wichita Falls, TX 76308-2099
Selected Paper prepared for presentation at the 2016 Agricultural & Applied Economics
Association Annual Meeting, Boston, Massachusetts, July 31-August 2
Copyright 2016 by P. Lynn Kennedy and Pablo A. Garcia-Fuentes. All rights reserved. Readers
may make verbatim copies of this document for non-commercial purposes by any means,
provided that this copyright notice appears on all such copies.
A supply and demand estimation of the United States high fructose corn syrup market
Abstract
This paper analyses the market of high fructose corn syrup (HFCS) in the United States. It
develops a system of demand and supply equations for both the HFCS and soft drinks markets.
The system of equations is estimated through Two Stage Least Squares methods. The results
show that soft drinks are the main driver of the growth of demand for HFCS. In addition,
negative news on HFCS (the association of HFCS consumption and obesity) has a negative
effect on the growth of demand for soft drinks, but soft drinks per capita advertising has a
positive effect on the growth of demand for soft drinks and more than offsets the effect of
negative news.
Key words: high fructose corn syrup, sugar, soft drinks, supply, demand.
Introduction
The United States is not only the second largest consumer of caloric sweeteners in the world,
which includes high fructose corn syrup (HFCS), but also number one producer of HFCS and
number five producer of sugar (Korves, 2011). In the United States, HFCS has replaced sugar as
sweetener in many final goods such as beverages and processed food products in the United
States. The substitution of HFCS for sugar in beverages and processed foods has been mainly
driven by differences in prices between HFCS and sugar. HFCS price has been lower than sugar
price, so beverage and processed food manufacturers have been able to either reduce costs or
keep costs low. During the early 1970s, there was a shock to the world supply of sugar that
affected the U.S. soft drinks industry, the whole sale refined sugar price increased from 12.4¢/lb.
in 1973 to 56¢/lb. in December 1974, so the Coca-Cola company started to replacing 25 percent
of the sucrose in its Fanta soft drink with HFCS-42 in 1974 (Bode, Empie, and Brenner, 2014).
Other companies followed Coca-Cola and started replacing between 25 percent and 50 percent of
sucrose with HFCS-42 (Forrestal, 1982). However, HFCS-42 was not the best substitute for
sucrose in soft drinks manufacturing, so HFCS manufacturers increased the concentration of
fructose in HFCS-42 to 55 percent to develop a new sweetener called HFCS-55 (Blanchard,
1992). Consequently, the new HFCS-55 was introduced to the market in 1978 (Corn Annual,
1996).
Another shock to the world supply of sugar that affected the U.S. soft drinks industry
happened in 1979 and the price of the U.S. wholesale refined sugar increased from 21¢/lb. in
1979 to 52¢/lb. in October 1980 (Bode, Empie, and Brenner, 2014). As a result, Coca-Cola and
Pepsi started replacing between 25 percent and 50 percent of sucrose with HFCS-55 in soft
drinks manufacturing in 1980, while 100 percent of sucrose was replaced with HFCS-55 in soft
drinks manufacturing in 1984 (Perdegrast, 1993). Regarding HFCS-42, it has become a
substitute for sugar in the production of baked products, but it has also been used as sweetener in
the production of beverages. Korves (2011) reports that the USDA estimated that 2.9 million
tons of HFCS-42 were used in 2010 in the production of food and beverages. Figure 1 shows per
capita consumption of HFCS and soft drinks for the period 1992:01-2013:12. Note that rises and
falls in soft drinks consumption are associated with rises and falls in consumption of HFCS
respectively.
Figure 1. Per capita consumption of HFCS and soft drinks.
Note. Authors’ calculation using data from USDA/ERS, Beverage Digest, Advertising Age, Business News, and
American Beverage Association.
The above discussion suggests that there are important relationships between the HFCS
market and the soft drinks and processed food markets. Regarding economic analysis of the
HFCS market, Carman (1982), based on a 25-percent market share ceiling for HFCS use, pointed
out this relationship and suggested that the potential use of HFCS in beverage products would be
100 percent and that the potential use of HFCS in processed food products would be 75 percent.
Evan and Davis (2002) report that soft drinks, as final goods that use HFCS, has the strongest
effect in a derived demand analysis for HFCS. However, these studies did not analyze the HFCS
and the soft drinks markets together. More recently, White (2014) reports that, of the total HFCS
and sugar delivered to the different food industries in the United States, the beverage sector uses
72 percent of HFCS and nine percent of sugar, and that processed foods sector uses 10 percent
HFCS and seven percent of sugar. Given this, it is likely that shocks to the soft drinks market or
the processed food market can affect the HFCS market.
To our knowledge no previous research has analyzed the HFCS market and the soft
drinks market together. This research aims to add to the literature on the economics of the HFCS
market by using a simultaneous system of supply and demand equations for both the HFCS and
soft drinks. The main purpose of this research is to assess the effect of soft drinks, as final goods
that use HFCS, on the demand for HFCS and the effects of soft drinks advertising and negative
news about HFCS (HFCS consumption and obesity) on the demand for soft drinks. Negative
news on HFCS is a proxy for the relation between HFCS consumption and obesity. The results
show that soft drinks is the main driver of the demand for HFCS and that negative news on
HFCS has a negative and significant effect on the demand for soft drinks, but soft drinks
advertising more than offsets the effect of negative news.
The rest of the paper is organized as follows. The second section provides a review of the
literature, followed by a description of the methodology and data. The next section presents a
discussion of the results. The last section presents conclusions and suggestions for further
research.
Literature Review
Carman (1982) is one of the earlier studies that focused on the HFCS market and
developed a projection of the U.S. HFCS market for the period 1981-1990 based on a HFCS
ceiling market share, a logistic function, total caloric sweetener demand, and annual data from
1967 to 1980. Based on a 25-percent market share ceiling use of HFCS, Carman (1982)
suggested that the potential use of HFCS in beverage products would be 100 percent and that the
potential use of HFCS in processed food products would be 75 percent and identified these two
food categories as the first and second highest potential users of HFCS. Carman (1982) estimated
a sweetener demand equation to develop a projection of the HFCS market for the period 1981-
1990 that used 1980 as the based period, the 1975-80 average price for sweeteners, the median
growth of the population, an estimated annual per capita consumption for caloric sweetener of
128 pounds for the year 1980, and an annual increase in real income of 1.5 percent. The
projection showed that total demand for sweeteners and demand for HFCS depended on, mainly,
population growth. HFCS was projected to increase from 21.93 pounds per capita in 1981 to
33.15 pounds per capita in 1990.
Barros (1992) assessed the effect of sugar prices on the consumption of HFCS in the
United States. He developed a model of demand and supply of HFCS expressed in growth rates
and estimated a reduced form of the demand for HFCS that covered the period 1971-1988. His
main findings were that high sugar prices (such as those in 1970s and 1980) were not the only
factors that promoted growth in HFCS consumption, a division of market shares for HFCS and
sugar in the sweetener market was not strongly supported, and protected sugar prices were not
the only factor affecting the growth of HFCS consumption.
Evans and Davis (2002) studied the dynamics of the U.S. high fructose sweetener market
and developed a derived demand model for HFCS. Their model’s assumptions are that HFCS
price has been less than sugar price, HFCS behaves as sugar substitute only on a given range of
the sugar demand curve, and HFCS price is relatively inelastic. They derive a reaction function
in the spirit of the Stackelberg model in which HFCS suppliers are the sophisticated ones, while
sugar suppliers are the naive ones. The model suggests that given that sugar prices are higher
than HFCS price, and as long as HFCS price is above its associated average cost, HFCS
suppliers are able to make economic rents. The results show that demand for HFCS is relatively
less elastic and suggest that decreases in sugar price are not likely to change the pattern of
sweetener use in the United States. The results also show that the price of soft drinks was,
economically, the most important explanatory factor in the HFCS demand equation.
The above literature review shows that HFCS demand is related to final products that use
HFCS. Korves (2011) reports that the USDA estimated that, in 2010, 2.9 million tons of HFCS-
42 were used in the production of food and beverages while 4.6 million tons of HFCS-55 were
used in the production of soft drinks. This suggests that the derived demand for HFCS is mainly
driven by the demand for processed foods and soft drinks. Therefore, it is important to conduct
an economic analysis of the U.S. HFCS market that accounts for the behavior of the soft drinks
market.
Methodology and Data
Model
The model is a system of supply and demand equations for the HFCS and soft drinks
markets. It is estimated using two-stage least squares (2SLS). In the case of the HFCS market,
we have
Demand: 𝑄ℎ𝑓𝑐𝑠 = 𝑓(𝑃ℎ𝑓𝑐𝑠, 𝑃𝑠, 𝐼𝑛𝑐, 𝑄𝑠𝑑) (1)
where 𝑄ℎ𝑓𝑐𝑠 is per capita quantity of HFCS, 𝑃ℎ𝑓𝑐𝑠 is the real list price of HFCS in cents per
pound, 𝑃𝑠 is real U.S. wholesale price of refined beet sugar in cents per pound, 𝐼𝑛𝑐 is real
personal disposable income, and 𝑄𝑠𝑑 is per capita quantity of soft drinks.
Supply: 𝑄ℎ𝑓𝑐𝑠 = 𝑓(𝑃ℎ𝑓𝑐𝑠, 𝑃𝑐𝑜 , 𝑅𝑖𝑟, 𝑃𝑒𝑡, 𝑃𝑒𝑙, 𝑃𝑐𝑛, 𝑇𝑟𝑒𝑛𝑑 ) (2)
where 𝑃𝑐𝑜 is real U.S. price of corn oil in cents per pound, 𝑅𝑖𝑟 is real interest rate, 𝑃𝑒𝑡 is real
price of ethanol, 𝑃𝑒𝑙 is real price of electricity in cents per kilowatt hour, 𝑃𝑐𝑛 is real price of
corn, and 𝑇𝑟𝑒𝑛𝑑 is used to control for changes in technology. Equation (1) includes per capita
quantity of soft drinks as a link between the HFCS market and the soft drinks market.
In the case of the Soft Drinks market, we have
Demand: 𝑄𝑠𝑑 = 𝑓(𝑃𝑠𝑑 , 𝐼𝑛𝑐, 𝐷𝑛𝑛, 𝐴𝑑𝑣𝑠𝑑 , 𝑇𝑟𝑒𝑛𝑑) (3)
where 𝑄𝑠𝑑 is per capita quantity of soft drinks, 𝑃𝑠𝑑 is real soft drinks producer price index, 𝐼𝑛𝑐 is
real personal disposable income, 𝐷𝑛𝑛 is a dummy variable for negative news about HFCS that
takes the value of one from January, 2004 to December, 2013 and zero otherwise, 𝐴𝑑𝑣𝑠𝑑 is per
capita soft drinks advertising expenses, and 𝑇𝑟𝑒𝑛𝑑 is used to control for changes in consumer
tastes and preferences.
Supply: 𝑄𝑠𝑑 = 𝑓(𝑃𝑠𝑑, 𝑃𝑒𝑙, 𝑇𝑟𝑒𝑛𝑑) (4)
where 𝑃𝑒𝑙 is real price of electricity in cents per kilowatt hour, 𝑃ℎ𝑓𝑐𝑠 is the real list price of HFCS
in cents per pound, and 𝑇𝑟𝑒𝑛𝑑 is used to control for changes in technology.
Data
This research uses monthly data that covers the period from January 1992 to December
2013. Although beverage processors began to use HFCS in Soft Drinks in the early 1970s,
production and disappearance data was not available until beginning in 1992. Given that the
concerns with HFCS consumption did not become apparent until the early to mid-2000s, the
period from 1992 through 2013 provides a good sample on either side of 2004 when the Bray
Report was released.
Prices of HFCS-42 and HFCS-55 are list prices in cents per pound of dry weight obtained
from USDA/ERS. Per capita quantity of HFCS-42 and HFCS-55 in pounds was constructed by
interpolation using quarterly data from the USDA/ERS and Proc Expand in SAS. Price of corn is
in dollars per bushel and is obtained from the USDA/ERS. Price of electricity is cents per kWh is
from USEIA. Real interest rate is computed by subtracting Moody's Seasoned AAA Corporate
Bond Yield from inflation rate. Moody's Seasoned AAA Corporate Bond Yield data is obtained
from the Federal Reserve Bank of St. Louis. We seasonally adjusted this data using the procX12
procedure in SAS. Inflation is the percentage change of GDP deflator 2009. Monthly data on
GDP deflator was constructed by interpolation using seasonally adjusted data from the Bureau of
Economic Analysis (BEA) and Proc Expand in SAS. Wholesale price of refined beet sugar is in
cents per pound is obtained from the USDA/ERS. Soft drinks producer price index (PPI) is
obtained from the Bureau of Labor Statistics. Negative news is a dummy variable that takes the
value of 1 from January 2004 onwards and zero otherwise. It is based on the publication date of
the paper by Bray, Nielsen, and Popkin (2004) that suggests that overconsumption of HFCS in
beverages is likely to be associated with obesity. Personal disposable income is seasonally
adjusted data from the BEA. Per capita quantity of soft drinks is in gallons and was constructed
by interpolation using yearly data from USDA/ERS, Beverage Digest, Advertising Age,
Business News, and the American Beverage Association. Price of corn oil is in cents per pound
from the USDA, Agricultural Marketing Service, Monthly Feedstuff Prices and Milling and
Baking News. Price of ethanol is dollars per gallon obtained from USDA/ERS. Per capita
advertising expense of soft drinks is constructed by using data from COMPUSTAT. We obtained
U.S. sales and total sales for Coca-Cola, Pepsico, and Dr. Pepper from COMPUSTAT’s
historical segments and data on total advertising expenses for these companies from
COMPUSTAT Monthly Updates-Fundamental Annuals. We computed yearly advertising
expenses for Coca-Cola, Pepsico, and Dr. Pepper by multiplying each company’s share of U.S.
sales in total sales by each company’s total advertising expenses. We created monthly
advertising expenses data for each company through interpolation using Proc Expand in SAS.
Total advertising for soft drinks in the U.S. was created by adding up each company’s monthly
advertising expenses. Population is obtained from the BEA, Table 2.6. Personal Income and Its
Disposition, Monthly. All monetary figures are in 2009 dollars. Variable definitions, data
sources, and descriptive statistics are in Appendices 1 and 2, respectively
Results
The results of the 2SLS estimations that use time series data that cover the period
1992:01 to 2013:12 are shown in Tables 1 and 2. Given the results of the Augmented Dickey-
Fuller and Phillips-Perron stationarity tests, we use first differences of the logs of each series in
the 2SLS estimations. Thus, demand and supply are expressed as growth rates. Further, first
differences have been found to contribute to reduce serial correlation and to improve the
statistical properties of the estimates (George and King, 1971). Additionally, all series are
seasonally adjusted.
Table 1 shows results for the case of demand and supply estimations of HFCS-55 and
soft drinks. The main relation between these two markets is because HFCS-55 has become a
substitute for sugar in soft drinks. Each equation in the system is a double-log model, so the
estimates are growth rates. Note that the results of the system of equations correspond with the
laws of demand and supply. The most interesting results are given by the positive and significant
effect of the quantity of soft drinks on the demand for HFCS-55. This suggest that one
percentage point increase in the growth rate of soft drinks increases the growth rate of demand
for HFCS-55 by 1.26 percentage points. The results also suggest that quantity of soft drinks is
the main driver of the demand for HFCS-55.
The demand equation for soft drinks also shows interesting results. Per capita income is
negative and highly significant. This suggests that soft drinks is an inferior good and that as
income grows the growth rate of demand for soft drinks decreases. The estimate on negative
news about HFCS-55 has a negative and significant effect on the growth of demand for soft
drinks. This suggests that people’s concerns about consumption of HFCS and obesity is
decreasing the growth of demand for soft drinks. However, the estimate on the growth of soft
drinks per capita advertising has a positive and significant impact on the growth of demand for
soft drinks. This coefficient is about 15 times the coefficient on negative news and more than
offsets the effect of negative news. So, one percentage point increase in the growth rate of per
capita soft drinks advertising increases the growth of demand for soft drinks by 0.0167
percentage points a month after accounting for the effect of negative news. Further, a 10
percentage point increase in per capita advertising increases growth of demand for soft drinks by
0.1778 percentage points which increases growth of demand for HFCS-55 by 0.22 percentage
points per month.
The results of Table 1 have shown that the market for HFCS-55 depends on the soft
drinks market and that advertising has an important direct impact on the growth of demand for
soft drinks and an important indirect impact on the growth of demand for HFCS-55 through soft
drinks quantity. This finding is in line with Evans and Davis’ (2002) study that reports that soft
drinks had the strongest effect in a HFCS derived demand estimation. In this study, the price of
soft drinks was used as a proxy for final goods that contain HFCS. However, they did not
estimate the demand and supply of HFCS and soft drinks together. Additionally, Table 1 results
give support to Carman’s (1982) study which reported estimates of HFCS ceilings regarding its
use in different food products and suggested that beverages and processed foods were the food
products that would allow for the largest use of HFCS. Carman (1982) also developed a
projected consumption of HFCS for the period 1981-1990 and suggested that the growth in total
sweetener demand would depend on population growth.
Table 1. Two Stage Least Squares Regression Results for Demand and Supply of High Fructose
Corn Syrup-55 and Demand and Supply of Soft Drinks, 1992:01-2013:12
a. Demand and Supply of High Fructose Corn Syrup-55(HFCS-55)
Demand Supply
Variables
Dep. Var.: Per capita
quantity of HFCS-55
Variables Dep. Var.: Per
capita quantity of
HFCS-55
Constant 0.0003
(0.43)
Constant 0.0049*
(1.80)
Fd ln(real price of HFCS-55) -0.0044
(0.05)
Fd ln(real price of HFCS-
55)
0.1850
(0.57)
Fd ln(real price of wholesale
refined beet sugar)
0.0217
(1.20)
Fd ln (real price of corn
oil)
-0.0239
(-1.39)
Fd ln(real per capita personal
disposable income)
0.1506
(1.42)
Fd ln(real interest rate) -0.0179
(0.65)
Fd ln(per capita quantity of
soft drinks)
1.2609**
(2.21)
Fd ln(real price of ethanol) 0.0053
(0.26)
Fd ln(real price of
electricity)
-0.0812
(0.48)
Fd ln(real price of corn) 0.0044
(0.24)
Trend 0.0000*
(-1.91)
R-squared 0.0295 R-squared 0.0340
Adj. R-squared 0.0145 Adj. R-squared 0.0075
Obs. 263 Obs. 263
b. Demand and Supply of Carbonated Soft Drinks
Demand Supply
Variables Dep. Var.: Per capita
quantity of soft drinks
Variables Dep. Var.: Per
capita quantity of
soft drinks
Constant 0.0018***
(5.69)
Constant 0.0019***
(5.29)
Fd ln(soft drinks PPI) -0.0100
(-0.34)
Fd ln(soft drinks PPI) 0.0037
(0.03)
Fd ln(real personal
disposable income)
-0.0898***
(-5.58)
Fd ln(real price of
electricity)
-0.0069
(0.21)
Dummy variable for negative
news on HFCS
-0.0012**
(-2.22)
Trend -0.0000***
(-9.35)
Fd ln(real per capita
advertising for soft drinks)
0.0179**
(2.07)
Trend -0.0000***
(3.29)
R-squared 0.3522 R-squared 0.2582
Adj. R-squared 0.3396 Adj. R-squared 0.2496
Obs. 263 Obs. 263
Note: Asterisks indicate significance at the 10 percent (*), 5 percent (**), and 1 percent (***) level respectively.
Values in parenthesis are t-values. Fd means first difference. Ln is the natural logarithm operator.
Table 2 shows results for the case of demand and supply estimations of HFCS-42 and
soft drinks. HFCS-42 has become a substitute for sugar in the production of baked products, but
it has also been used as sweetener in the production of beverages. Korves (2011) reports that the
USDA estimated that 2.9 million tons of HFCS-42 were used in 2010 in the production of food
and beverages. Each equation in the system is a double-log model, so the estimates are growth
rates. Note that the results of the system of equations correspond with the laws of demand and
supply. In this estimation system, quantity of soft drinks is a proxy for food and beverages that
use HFCS-42 as sweetener. The most interesting results are given by the positive and significant
effect of the quantity of soft drinks on the demand for HFCS-42. This suggest that one
percentage point increase in the growth rate of soft drinks increases the growth rate of demand
for HFCS-42 by 1.34 percentage points. The results also suggests that quantity of food and
beverages products is the main driver of the demand for HFCS-42.
The demand equation for soft drinks also shows interesting results. Per capita income is
negative and highly significant. This suggests that soft drinks is an inferior good and that as
income grows the growth rate of demand for soft drinks decreases. The estimate on negative
news about HFCS-55 has a negative and significant effect on the growth of demand for soft
drinks. This suggests that people’s concerns about consumption of HFCS and obesity is
decreasing the growth of demand for soft drinks. However, the estimate on the growth of soft
drinks per capita advertising has a positive and significant impact on the growth of demand for
soft drinks. This coefficient is about 15 times the coefficient on negative news and more than
offsets the effect of negative news. So, one percentage point increase in the growth rate of per
capita soft drinks advertising increases the growth of demand for soft drinks by 0.0167
percentage points a month after accounting for the effect of negative news. Further, a 10
percentage point increase in per capita advertising increases growth of demand for soft drinks by
0.1778 percentage points which increases growth of demand for HFCS-42 by 0.24 percentage
points per month.
The results of Table 2 have shown that the market for HFCS-42 depends on the food
products that use HFCS-42 as sweetener and that advertising has an important direct impact on
the growth of demand for soft drinks and an important indirect impact on the growth of demand
for HFCS-42 through soft drinks quantity. In this case soft drinks quantity is a proxy for final
products that use HFCS-42 as sweetener. This finding is in line with Evans and Davis’ (2002)
study that reports that soft drinks had the strongest effect in a HFCS derived demand estimation.
Again, Table 2 results give support to Carman’s (1982) study that suggested that beverages and
processed foods were the food products that would allow for the largest use of HFCS and that the
growth in total sweetener demand would depend on population growth.
Table 2. Two Stage Least Squares Regression Results for Demand and Supply of High Fructose
Corn Syrup-42 and Demand and Supply of Soft Drinks, 1992:01-2013:12
a. Demand and Supply of High Fructose Corn Syrup-42(HFCS-42)
Demand Supply
Variables
Dep. Var.: Per capita
quantity of HFCS-42
Variables Dep. Var.: Per capita
quantity of HFCS-
42
Constant -0.0004
(-0.52)
Constant 0.0042
(0.88)
Fd ln(real price of HFCS-42) -0.0032
(0.02)
Fd ln(real price of HFCS-
42)
0.1483
(0.24)
Fd ln(real price of wholesale
refined beet sugar)
0.0202
(1.07)
Fd ln (real price of corn
oil)
-0.0230
(-1.15)
Fd ln(real personal
disposable income)
0.1864
(1.60)
Fd ln(real interest rate) -0.0119
(0.43)
Fd ln(per capita quantity of
soft drinks)
1.3426*
(1.96)
Fd ln(real price of
electricity)
-0.0258
(0.13)
Fd ln(real price of corn) 0.0041
(0.14)
Trend -0.0000
(1.07)
R-squared 0.0298 R-squared 0.0334
Adj. R-squared 0.0148 Adj. R-squared 0.0107
Obs. 263 Obs. 263
b. Demand and Supply of Carbonated Soft Drinks
Demand Supply
Variables Dep. Var.: Per capita
quantity of soft drinks
Variables Dep. Var.: Per capita
quantity of soft
drinks
Constant 0.0018***
(5.69)
Constant 0.0018***
(3.99)
Fd ln(soft drinks PPI) -0.0093
(0.31)
Fd ln(soft drinks PPI) 0.0421
(0.25)
Fd ln(real personal
disposable income)
-0.0898***
(5.58)
Fd ln(real price of
electricity)
-0.0184
(0.36)
Dummy variable for negative
news on HFCS
-0.0012**
(-2.22)
Trend -0.0000***
(9.12)
Fd ln(real per capita
advertising for soft drinks)
0.0179 **
(2.07)
Trend 0.0000***
(-3.29)
R-squared 0.3521 R-squared 0.2542
Adj. R-squared 0.3396 Adj. R-squared 0.2456
Obs. 263 Obs. 263
Note: Asterisks indicate significance at the 10 percent (*), 5 percent (**), and 1 percent (***) level respectively.
Values in parenthesis are t-values. Fd means first difference. Ln is the natural logarithm operator.
Conclusion
This study uses monthly data that covers the period 1992:01-2013:12. It develops a
system of demand and supply models to explain the relation between the HFCS market and the
soft drinks market in the United States. This framework identifies important drivers of the
demand for soft drinks and the effect of soft drinks on the demand for HFCS. The system of
demand and supply equations is estimated by using first differences of all series and 2SLS
methods.
The results show that soft drinks, as final goods that use HFCS as input, is the most
important factor that drives the growth of demand for HFCS in the United States. This is in line
with past research that finds soft drinks as the most important driver in a derived demand
estimation of HFCS (Evans and Davis, 2002). In addition, regarding the demand for soft drinks,
negative news about the relation between HFCS consumption and obesity decreases the growth
of demand for soft drinks. However, the coefficient on soft drinks advertising is about 15 times
the coefficient on negative news, so it more than offsets the effect of negative news on HFCS
and increases the growth of demand for soft drinks. So, advertising has an important direct
impact on the growth of demand for soft drinks and an important indirect impact on the growth
of demand for HFCS through soft drinks quantity. Therefore, the results suggest that the market
for HFCS depends on the market for soft drinks.
Regarding future research, we plan to expand the period covered by the study back to
1984. This is the year when Coca-Cola and Pepsi started a 100 percent use of HFCS-55
(Pendergrast, 1993). We also plan to develop a dynamic simultaneous equations model to deal
with the time series aspects of the data.
References
Barros, Alexandre Rands. 1992. "Sugar prices and high‐fructose corn syrup consumption in the
UNITED STATES." Journal of Agricultural Economics 43(1): 64-73.
Bode, John W., Mark W. Empie, and Kyd D. Brenner. 2014. Evolution of High Fructose Corn
Syrup within the Sweeteners Industry. In Fructose, High Fructose Corn Syrup, Sucrose and
Health, ed. James M. Rippe, 137-148, New York: Springer Science+Business Media.
Blanchard PH. 1992. Technology of corn wet milling and associated processes. London:
Elsevier.
Bray, George A., Samara Joy Nielsen, and Barry M Popkin. 2004. "Consumption of high-
fructose corn syrup in beverages may play a role in the epidemic of obesity." The American
journal of clinical nutrition 79(4): 537-543.
Carman, Hoy F. 1982. “A Trend Projection of High Fructose corn Syrup Substitution for Sugar.”
American Journal of Agricultural Economics, 64(4):625-633.
Corn Refiners Association.1996. Corn Annual. Washington, DC: Corn Refiners Association.
Evans, Edward A. and Carlton G. Davis. 2002. Dynamics of the United States High Fructose
Corn Sweetener Market. In Sugar and Related Sweetener Markets: International
Perspectives, ed. Andrew Schmitz, Thomas H. Spreen, William A. Messina Jr., and Charles
B. Moss, 281-297, New York: CABI Publishing.
Forrestal D. J. 1982. The Kernel and the Bean. New York, NY: Simon and Schuster.
George, Poykayil Simon, and Gordon A. King. 1971. Consumer demand for food commodities in
the United States with projections for 1980. Berkeley: University of California, Berkeley.
Korves, Ross. 2011. "High fructose corn syrup (HFCS) in the US caloric sweetener supply."
Paper prepared for the Corn Refiners Association. Accessed at
https://scholar.google.com/scholar?q=+High+Fructose+Corn+Syrup+%28HFCS%29+in+the
+U.S.+Caloric+Sweetener+Supply&btnG=&hl=en&as_sdt=0%2C19.
Pendergrast M. 1993. For God, country and Coca-Cola. New York: Charles Scribner and Sons.
White, John S. 2014. Sucrose, HFCS, and Fructose: History, Manufacture, Composition,
Applications, and Production. In Fructose, High Fructose Corn Syrup, Sucrose and Health,
ed. James M. Rippe, 13-33, New York: Springer Science+Business Media.
Appendix 1 Variable definitions and data sources
Variable name Variable definition Source
Price of
HFCS-42
Nominal list price in cents per pound
of dry weight
USDA/ERS, Sugar and Sweeteners
Yearbook Tables, Table 9. Accessed
at http://www.ers.usda.gov/data-
products/sugar-and-sweeteners-
yearbook-tables.aspx
Price of
HFCS-52
Nominal list price in cents per pound
of dry weight
USDA/ERS, Sugar and Sweeteners
Yearbook Tables, Table 9. Accessed
at http://www.ers.usda.gov/data-
products/sugar-and-sweeteners-
yearbook-tables.aspx
Quantity of
HFCS-42
Total deliveries of HFCS-42 in tons USDA/ERS, Sugar and Sweeteners
Yearbook Tables, Table 30. Accessed
at http://www.ers.usda.gov/data-
products/sugar-and-sweeteners-
yearbook-tables.aspx
Quantity of
HFCS-55
Total deliveries of HFCS-55 in tons USDA/ERS, Sugar and Sweeteners
Yearbook Tables, Table 30. Accessed
at http://www.ers.usda.gov/data-
products/sugar-and-sweeteners-
yearbook-tables.aspx
Price of corn Nominal price in dollars per bushel USDA/ERS, Feed Grains Database,
Custom Query for No. 2 Yellow Corn
Market Prices, Central Illinois.
Accessed at
http://www.ers.usda.gov/data-
products/feed-grains-database/feed-
grains-custom-query.aspx
Price of
electricity
Nominal price in cents per kWh United States Energy Information
Administration. http://www.eia.gov
Real interest
rate
Real interest rate. It is computed by
subtracting Moody's Seasoned AAA
Corporate Bond Yield from inflation
rate.
Own calculations.
Moody's
Seasoned
AAA
Corporate
Bond Yield
Moody's Seasoned AAA Corporate
Bond Yield
Federal Reserve Bank of St. Louis
Inflation Monthly Percentage change of GDP
deflator. It was constructed by
interpolation using quarterly data
and SAS proc expand.
Own calculations.
GDP deflator Seasonally adjusted quarterly GDP
deflator.
Bureau of Economic Analysis
Appendix 1 (continued)
Variable name Variable definition Source
Price of refined
beet sugar
Nominal wholesale price of refined
beet sugar in cents per pound.
USDA/ERS, Sugar and Sweeteners
Yearbook Tables, Table 5. Accessed
at http://www.ers.usda.gov/data-
products/sugar-and-sweeteners-
yearbook-tables.aspx
Soft drinks
price
Soft drinks producer price index Bureau of Labor Statistics
Negative news
on HFCS
Dummy variable that takes the value
of 1 from January 2004 onwards and
zero otherwise. It is based on the
publication date of the paper by Bray,
Nielsen, and Popkin (2004).
Own calculations.
Per capita
personal
disposable
income
Seasonally adjusted per capita
personal disposable income.
Bureau of Economic Analysis
Monthly per
capita quantity
of soft drinks
Monthly per capita quantity of soft
drinks in gallons. It was constructed
by interpolation using yearly data.
Own calculations
Annual per
capita quantity
of soft drinks
Annual per capita quantity of soft
drinks in gallons.
USDA/ERS, Beverage Digest,
Advertising Age, Business News,
and American Beverage
Association.
Price of corn
oil
Nominal price of corn oil in cents per
pound.
USDA, Agricultural Marketing
Service, Monthly Feedstuff Prices
and Milling and Baking News.
Price of
ethanol
Price of ethanol in dollars per gallon. USDA/ERS, U.S. Bioenergy
Statistics, Table 14, Fuel ethanol,
corn and gasoline prices, by month.
Accessed at
http://www.ers.usda.gov/datafiles/U
S_Bioenergy/Prices/table14.xls
Per capita
advertising of
soft drinks
Per capita advertising expense of soft
drinks in cents. It is constructed using
U.S. sales, total sales, and total
advertising expenses data for Coca-
Cola, Pepsico, and Dr. Pepper.
Own calculations.
U.S. sales of
soft drinks
U.S. sales for Coca-Cola, Pepsico,
and Dr. Pepper.
COMPUSTAT’s historical segments
Soft drinks
advertising
expense
Total advertising expenses for Coca-
Cola, Pepsico, and Dr. Pepper.
COMPUSTAT Monthly Updates-
Fundamental Annuals.
Population Population Bureau of Economic Analysis.
Appendix 2 Summary statistics, monthly values for the period 1992:1-2013:12 Variable N Mean Std. Dev Minimum Maximum
HFCS-55 per capita quantity (pounds) 264 2.90 0.30 2.37 3.45
Real HFCS-55 price 264 26.14 5.54 16.64 37.18
HFCS-42 per capita quantity (pounds) 264 1.92 0.22 1.40 2.21
Real HFCS-42 price 264 23.34 5.87 12.44 34.43
Per capita quantity of soft drink (gallons) 264 4.13 0.29 3.54 4.50
Soft drink price index 2009 264 84.37 12.63 65.12 107.23
Real wholesale refined sugar price 264 34.35 7.96 23.61 57.30
Real corn oil price 264 35.61 12.44 12.08 84.06
Real interest rate 264 6.04 1.27 3.21 8.46
Real ethanol price 264 1.86 0.47 1.13 3.70
Real electricity price 264 6.24 0.43 5.31 7.19
Real corn price 264 3.58 1.41 1.84 7.44
Real personal disposable per capita income 264 32,237.65 3,744.31 25,886 38,175
Per capita soft drink advertising (cents) 264 0.62 0.08 0.50 0.74