|
The Impact of Big-Box Stores on Retail Food Prices and the
Consumer Price Index
Ephraim Leibtag
Economic Research Report No. (ERR-33), December 2006
Nontraditional retailers such as Wal-Mart, Costco, and Target
have gained more of the consumer food dollar over the past 10
years. The share of sales going to traditional retailers (conventional
supermarkets, superstores, and food-drug combination stores)
fell from 82 percent in 1998 to 69 percent in 2003.
What Is the Issue?
Over the past 20 years, annual food price changes, as measured
by the Consumer Price Index (CPI), have averaged 3 percent per
year. Meanwhile, food prices for the same item can vary by more
than 10 percent from one type of store to another in a given
year. The CPI measure of food price inflation is based on a
sample of food items selected from a sample of retail food outlets,
and the selection of stores has not been updated quickly enough
to reflect the volume of food now sold through big-box stores.
Since the current CPI for food does not fully account for the
lower prices offered by these nontraditional retailers, including
prices from all store formats would likely indicate a lower
rate of price inflation than the CPI estimate.
What Did the Study Find?
Previous studies have demonstrated that food prices at nontraditional
retailers are 8-27 percent lower than at large supermarket chains.
However, these comparisons across store formats did not account
for quality or package size differences for some food products.
To address these concerns, comparisons in this report are for
similar package sizes and more specifically defined food items:
namely, dairy products and eggs.
Even when controlling for similar-sized packages, dairy prices
are 5 to 25 percent lower at nontraditional retailers than at
traditional supermarkets. For example, skim and low-fat milk
prices are consistently 5-12 percent lower at nontraditional
stores. Even more price variation exists in random-weight cheese
products; a pound of Swiss cheese averaged $4.71 at grocery
stores in 2003, but just $3.77 at nontraditional retailers and
mass merchandisers.
Since food-at-home inflation averaged 2.2 percent per year
over 1998-2003, a discrepancy of 5-25 percent in price between
store formats is relatively large. If the difference in food
prices across store formats applies to many food categories,
the official estimates of price changes might be overstating
the actual rate of change. If that is the case, another way
to estimate price change might be to track consumer purchase
behavior and adjust observed price changes in a given category.
This could be accomplished using an expenditure-weighted measure
of price change, with frequently updated measures from scanner
data sources.
How Was the Study Conducted?
This study uses ACNielsen Fresh Foods Homescan scanner panel
data for 1998-2003. The annual data are from a consumer panel
consisting of about 8,000 representative households across the
United States; the data included purchase as well as demographic
information. Panelists recorded both their UPC-coded transactions
and their random-weight (non- UPC) food purchases over the year(s)
that they participate in the survey.
Average prices were calculated for a wide variety of dairy
products (24 fixed-weight and 8 random-weight) commonly purchased
at both traditional and nontraditional stores. The average prices
were calculated by taking the total weighted expenditures for
a given product and dividing by the total weighted quantity
that was purchased. These average prices were then weighted
using the projection factors for each household in the sample
to arrive at a national average in each food category. The average
prices were then used to calculate an average annual price change
for each food category to compare with price change in the corresponding
CPI categories.
|
