Competing with the Discount
A Study by
Dr. Kenneth E. Stone
Professor of Economics &
Iowa State University
IMPACT OF WAL-MART STORES AND OTHER MASS
MERCHANDISERS IN IOWA, 1983-1993*
The nature of retailing has changed dramatically in the last decade,
compared to previous decades. In the last decade there has been a great
proliferation of discount general merchandise stores as Wal-Mart, K Mart, Target
and several regional chains. In addition there has been a great expansion of
membership warehouse clubs, such as Sam’s, Pace, Cosco, and Price Club. There
has also been rapid expansion of “category killer” stores such as Home Depot,
Circuit City, Best Buy, Toys “R” Us and others. These stores are called category
killer stores because they have a very large selection within a narrow category of
merchandise, along with low prices, and they “ kill” smaller local stores within the
same “category”. We have also seen the development of many new factory outlet
malls and the spread of specialty mail order. The net result of this expansion is the
saturation of many retail markets, or what is commonly referred to as the
overstoring of America. Many retail markets have more retail stores than can
possibly be supported, and it would appear that a major shakeout is coming in the
Why the Study?
This study is an annual update of my original study of the impact of Wal-
Mart stores completed in 1988. The 1988 study was conducted at the request of
several Iowa merchants and chamber of commerce executives. Wal-Mart stores
began locating in Iowa in 1983 and local residents did not know much about them,
thus prompting requests for information. Iowa has an excellent sales tax reporting
system, but because of a lag in reporting and in the initial slowness of the Wal-
Mart expansion, it was 1988 before enough data was available to conduct the first
The purpose of the studies was to document changes in retail sales in the
host community and in surrounding communities. Once the changes were
identified, then educational programs could be designed to assist local business
people better cope or co-exist in the new mass merchandiser environment.
*by Kenneth E. Stone, Professor of Economics, Iowa State University
Initially, it was decided to study the impact of Wal-Mart stores in
communities between 5,000 and 30,000 population, because in 1988 that was
where most of the stores were located. Since the original study, Wal-Mart stores
have located in all the larger cities, but the impact there will be the subject of a
separate study. This update through fiscal year 1993 will continue to examine the
impact on communities with populations between 5,000 and 30,000 population.
Retail sales data came from the Iowa Retail Sales and Use Tax Reports,
published annually by the Iowa Department of Revenue and Finance. The
department has a long history of publishing this data for virtually every town and
city in the state. Sales of all taxable goods and services are reported by two digit
Standard Industrial Classification (SIC) Codes for communities over 2,500
population, if the community has five or more businesses within the category. The
two digit classification is a broad description of businesses such as food, general
merchandise, building materials, etc. Statewide sales are also published at the
more detailed three digit SIC level. For example, in the building materials
category, statewide sales would be reported for hardware stores, lumber yards,
paint and glass stores, etc.
Comparison of Data.
The simplest method of analysis would have been to compare retail sales
for each of the communities. However, this is not a very satisfactory method since
it does not take into account changes in price inflation, population and economic
cycles. Instead, we chose to use a derivative called the pull factor in this
comparison. The pull factor is merely the per capita sales of the community
divided by statewide per capita sales. Community per capita sales are calculated
by dividing the community’s retail sales by the community population. Likewise,
the statewide per capita sales are computed by dividing the total state retail sales
by the state population. The equation for determining a pull factor is shown
PF = --------
Where: PF = pullfactor
PCSc = per capita sales for a community
PCSs = per capita sales for the state
For example if a town had a per capita sales of $8,500 per year and the
statewide per capita sales were also $8,500 per year, the pull factor for the town
would be calculated by dividing the community’s $8,500 by the statewide $8,500
and the pull factor would equal 1.0. The interpretation would be that the
community’s retail sales amounted to selling to the equivalent of 100 percent of
the community population in terms of full time customer equivalents. When a pull
factor is less than one (say 0.6, for example) the interpretation is that the
community is selling only to the equivalent of 60% of the community population.
Conversely, when a pull factor is greater than one (say 1.5, for example), it would
be concluded that the community is selling to the equivalent of 150 percent of the
community population, a fair sized trade area.
In fact, in most cases, some community residents would purchase some or
all of their retail goods in other communities, while other people from the
surrounding area would purchase some or all of their retail goods in the subject
community. As was stated before, the pull factor makes adjustments for changes
in population, price inflation and economic cycles. Pull factors can be computed
for total sales and for two digit SIC codes.
Comparison by Time Period.
The first Wal-Mart stores opened in Iowa in 1983, and then a small number
were opened each year thereafter, reaching a total of 45 by fiscal year 1993. This
study analyzes 34 towns with populations between 5,000 and 30,000 population.
We thought it important that we be able to measures the change in sales for
different types of stores, both in the host community and other communities for
successive years after a Wal-Mart opening and then to generalize as to the average
impacts after one year, two years, three years, etc. However, with the store
openings spread over such a long period of time, a method had to be devised to
make equitable and meaningful comparisons. That method is discussed below.
Base Year. The base year for comparison for each community was the last full
year before a Wal-Mart store opened there. For example, if a store opened in
Marshalltown in 1983, the base year was 1982. Each successive year of operation
would then be compared to the base year of 1982, using the pull factors. For
example, if Marshalltown’s pull factor for total sales in 1982 was 1.50 and in 1983
after one year of Wal-Mart it increased to 1.56, an increase of four percent was
calculated. The pull factor for 1984 would then be computed and compared to that
of 1982 to see the cumulative change after two years, etc. In other words,
successive year changes vary among towns, but in each year’s comparison,
changes in price inflation, population, and the state’s economy are always taken
into consideration. The first year changes for different communities are then
averaged as are the successive year changes, even though they may be for different
A Basic Premise. A basic premise lies at the heart of this study. The premise is
that in areas of somewhat static population (such as in states like Iowa) the size of
the retail “pie” is relatively fixed in size for a given geographical area.
Consequently, when a well-known national chain like Wal-Mart opens a large
store in a comparatively small town, it invariably will capture a substantial slice of
the retail pie. The end result is that other merchants in the area will have to make
do with smaller slices of the retail pie, or get out of business. In areas of the
country where the population is growing rapidly, there is room for more retail
establishments and the effect will be diluted considerably.
The study found both plusses and minuses for the merchants in the host
town. The major plus for most businesses was that in virtually all cases, total sales
in the town increased at a rate greater than average for the state. Apparently Wal-
Mart stores attracted customers into town from a greater radius than had occurred
before their entry into the town. After the first two or three years, however, most
Wal-Mart towns reached a peak in sales and then began a decline, with about 25
percent declining below the pre-Wal-Mart level. This decline appears to be
caused by a saturation of Wal-Mart stores and other competing stores. Two
simple rules of thumb explain the winners and the losers among host town
Rule 1. Merchants selling goods or services different from what Wal-Mart sells
become natural beneficiaries. In other words, since they are not competing
directly, many of them benefit from the spillover of the extra customers being
pulled into town by Wal-Mart.
Rule 2. Merchants selling the same goods as Wal-Mart are in jeopardy. In other
words, they are subject to losing some trade to Wal-Mart unless they change their
way of business.
Retail sales for non-Wal-Mart towns declined in all categories except food
stores after the opening of Wal-Mart stores. The magnitude of the sales declines
for non-Wal-Mart towns was found to be much greater in this study than it was in
earlier studies. The probable reason for the larger decline in sales is that the
density of Wal-Mart stores has increased substantially. In earlier studies, the few
Wal-Mart stores were widely scattered and residents of the non-Wal-Mart towns
were sometimes 50 or more miles from Wal-Mart towns, a disincentive for
traveling to shop. At the time of this study, few non-Wal-Mart towns were more
than 25 miles away from a Wal-Mart store, and apparently more people were
traveling to the Wal-Mart towns to shop.
Smaller outlying towns (populations below 5,000) appear to have borne the
brunt of all the new Wal-Mart stores and others. For example towns with
populations between 500 and 1,000 lost nearly 47 percent of their sales from 1983
to 1993. The results of the 1993 study are discussed below.
General merchandise stores consist of department stores and variety stores
and include discount general merchandise stores such as Wal-Mart, K Mart, and
Wal-Mart Towns. Figure 1 shows that Wal-Mart towns experienced a huge
increase in sales in the general merchandise category in the years following the
opening of a Wal-Mart store. On average, the sales after the first year increased
by 53.6 percent. Obviously, most of this increase in sales goes to the Wal-Mart
store. It is interesting to note, however, that the large initial increase in sales is not
maintained and sales increases over the base year drop from 53.6 percent the first
year to 43.6 percent in years three and five, respectively. This is probably due to a
curiosity factor the first year that brings in casual shoppers, as well as a steady
saturation of general merchandise stores that dilutes the sales of all.
Non-Wal-Mart Towns. The non-Wal-Mart towns are towns between 5,000 and
30,000 population that do not have a Wal-Mart store. Typically they have a
K Mart store or a regional discount store such as Pamida, Alco or Places.
General Merchandise Changes
*Changes are cumulative from base year.
As shown in Figure 1 general merchandise stores in non-Wal-Mart towns
did not fare well after Wal-Mart stores came into the area. On average, sales
declined by 5.2 percent after the first year and slipped to a 12.9 percent cumulative
decrease after five years. It appears that the Wal-Mart stores captured sales from
the general merchandise stores in the non-Wal-Mart towns.
Cities. In this study, cities were defined as municipalities with more than 50,000
population. In Iowa, there are eight cities over 50,000 population, with Des
Moines being the largest at approximately 200,000 population. It was surprising
to find that general merchandise sales declined in the cities after Wal-Mart stores
came into the state, since they had been increasing up until that time. Figure 1
indicates a decline of 2.8 percent after one year, but that deteriorated to 9.5 percent
cumulative decrease after five years. One explanation for the decrease of general
merchandise sales in the cities is that as Wal-Mart stores increased in numbers,
more and more local residents stayed at home to shop, rather than traveling to
larger cities. Another possible explanation is that the category killer stores in the
cities captured some of the sales that previously were made in the general
The home furnishings category is made up of furniture stores, major
appliance stores, consumer electronics stores, floor covering stores and
miscellaneous others such as drapery stores. Note that very little of this
merchandise is sold in a Wal-Mart store.
Wal-Mart Towns. Home furnishings stores in Wal-Mart towns fared well after
the opening of the Wal-Mart store. Figure 2 shows that after a first year setback of
2.3 percent, home furnishings sales increased by 3.2 percent over the base year
after three years and enjoyed an 8.7 percent cumulative increase after five years.
Anecdotal evidence suggests that towns with good home furnishings stores
experience an increase in this category because they benefit from the spillover
of additional shoppers drawn to town by the Wal-Mart store.
Non-Wal-Mart Towns. Home furnishings stores in towns without a Wal-Mart
store suffered substantial reductions in sales in the years after Wal-Mart’s opening.
Figure 2 shows that sales declined by 4.6 percent after the first year, 13.1 percent
after the three years and 16.9 percent after five years. Apparently consumers are
drawn from these towns to Wal-Mart towns and larger cities to shop and they
choose to purchase some of their home furnishings in these towns also.
Home Furnishings Changes
*Changes are cumulative from base year.
Cities. The eight cities in the study showed an average increase in home
furnishings sales of 3.2 percent after one year of Wal-Mart stores. After three
years sales were up 4.2 percent, but they declined slightly to 4.0 percent after five
years. Although the percentages are relatively small, the dollar amounts are
Eating and Drinking Places
Eating and drinking places obviously are restaurants and bars. Restaurants
have been and continue to be a growth business because more and more people are
eating away from home.
Wal-Mart Towns. Restaurant sales in Wal-Mart towns grew faster than the state
average. Figure 3 shows that after one year, restaurant sales were up 3.2 percent,
then climbed to a gain of 4.8 percent after three years, before declining to a 3.2
percent gain after five years. Apparently the Wal-Mart stores were drawing
more people into town to shop and before going home, some consumed meals
Eat & Drink Changes
*Changes are cumulative from base year.
Non-Wal-Mart Towns. Non-Wal-Mart towns experienced a decline in restaurant
sales after Wal-Mart stores opened in the area. Figure 3 indicates that sales
declined 3.2 percent after one year, 5.6 percent after three years and 7.9 percent
after five years, in spite of the fact that restaurant sales were rising statewide.
Apparently residents of non-Wal-Mart towns were out-shopping in Wal-Mart
towns and larger cities and consuming meals there instead of in their
Cities. The cities in the study showed a continual increase in eating and drinking
sales. Figure 3 shows that they were up 0.5 percent after one year, up 1.9 percent
after three years and up 2.9 percent after five years. These figures would seem to
reflect the normal growth pattern of sales in cities and were probably not
affected one way or another by the opening of Wal-Mart stores around the
Apparel stores consist of men’s, women’s and children’s clothing stores as
well as shoe stores. Miscellaneous apparel and accessory stores such as western
wear would also be included in this category.
Wal-Mart Towns. Apparel store sales in the Wal-Mart towns declined by 7.9
percent after the first year of a Wal-Mart store, as shown in Figure 4. Sales were
down 12.9 percent after three years and declined further to 17.9 percent after five
years. These are large losses and probably indicates that many of the local apparel
stores were positioned at the low end of the market, thereby competing directly
with the Wal-Mart apparel departments. In a few towns where there was a good
mix of upscale apparel stores, sales increases were enjoyed by apparel stores
since they were not competing directly, but benefited from the additional
traffic generated by the Wal-Mart store.
Non-Wal-Mart Towns. As shown in Figure 4, apparel store sales in non-Wal-
Mart towns showed a steady decline over the first five years. The losses were 7.5
percent after one year, 10 percent after three years and 13.1 percent after five
years. Again one would have to conclude that the loss of sales to the apparel
stores was probably at the low end since that is primarily what discounters
like Wal-Mart handle.
Cities. Iowa cities showed gradual gains in apparel sales and appeared to be little
affected by the growth of Wal-Mart stores. As shown in Figure 4, sales were up
0.9 percent after one year, and increased to 2.0 and 2.1 percent after three and five
*Changes are cumulative from base year.