Friday, July 18, 2008

Wal-Mart: No Effect on Small Businesses Activity

Myth: Mega discount store Wal-Mart is a plague set upon small “mom-and-pop” businesses. The instant Wal-Mart moves into town, all small businesses are destroyed in its path, leaving downtowns barren and empty. According to Robert Reich, Wal-Mart turns “main streets into ghost towns by sucking business away from small retailers.”

Reality: The popular belief that Wal-Mart has a significant negative effect on the size of the mom-and-pop business sector of the United States economy is statistically unfounded. After examining a plethora of different measures of small business activity and growth, examining both time series and cross-section data, and employing different geographic levels of data and different econometric techniques, it can be firmly concluded that Wal-Mart has had no significant impact on the overall size and growth of U.S. small business activity.

Source: Cato Institute study "Has Wal-Mart Buried Mom and Pop?"

MP: In the chart above, the slope of the regression line is actually positive and significantly different from zero, which suggests that states with more Wal-Mart stores actually have significantly higher levels of five-to-nine-employee establishments.


At 7/18/2008 2:40 PM, Blogger Unknown said...

I know I'm a pain.

But, there is another regression line starting from the outlier in the top left and going down with negative slope. Of course with less significance but there is one.

I mean, the data is too scattered to show anything.

Fundamentally speaking, statistics fall under the problem of induction.

You are trying to deduce a general rule from specific premises. But the fact is that the conclusion is already built in some of the premises:

(1) Wall-Mart will not build more stores in less populated areas so the effect will not show there as pronounced.

(2) Although Wal_Mart builds more stores in more populated areas, there are also more small business that survive frome the many more that would have survived had Wal_Mart not being there.

You also have selection bias. A few more data on the lower left and you get a reversal. The correct way to do this, IMO, would be to find places with NO Wal-Mart stores (maybe impossible) and compare to places with stores.

Otherwise, your conclusions are implicit in the premises and no deduction is possible.

Statistics cannot be used as a weapon against common sense.

However, in some cases, large department stores bring more market efficience, it is true, due to local small cartels that fix prices.

At 7/18/2008 2:58 PM, Blogger Unknown said...

Protagoras, my master, just read my post and said to me:

"stupid, you could have just said that a small positive slope does not imply the conclusion"

At 7/18/2008 3:05 PM, Anonymous Anonymous said...

"The correct way to do this, IMO, would be to find places with NO Wal-Mart stores (maybe impossible) and compare to places with stores."

Read the paper. that's what they did...compared the 5 states with the most Wal-Marts and the 5 with the fewest.

Another interesting fact the Professor didn't mention, is that the additional small business growth that occurs with Wal-mart stores appears be better businesses...the revenue and income of sole proprieterships also show an increase with the increase in the number of Wal-mart stores in the area.

At 7/18/2008 3:10 PM, Anonymous Anonymous said...

ETA: sorry, sophist...just realized the graph itself says "48 states" the paper they explain they started with the 5 states with highest and 5 states lowest # of Wal-marts per capita, and then moved on to find similar results among the other 48 states.

At 7/18/2008 3:15 PM, Blogger Matt S said...

I will say that I speak from personal experience when I say that big stores do hurt small businesses, at least in rural areas.
The emergence of a few Atlantic Superstores in Prince Edward Island, Canada (a province that can see its population double when it hosts an aerosmith concert), has been a negative for mom and pop grocers and general stores.
My uncle runs a gas bar/general store/ mini grocery with fresh produce and customizable butchered meats/ ice cream stand, which has been around since my grandfather Amos Gallant founded it about 76 years ago. However, the big box stores have cut down on revenues of stores like my uncle's and others.
Now, I think it's an absolute shame, because that general store has more history and cultural significance to more people than any chain store ever will, but at the same time I understand that the "market" works that way. the store might have to close or sell to another franchiser, because too many people are stopping by the "cheaper" big stores while in the main town instead of getting their groceries closer to home.

At 7/18/2008 3:16 PM, Anonymous Anonymous said...

I suspect this is correct, purely from my personal anecdotal observations those that are hit hardest when Wal-Mart enters the area are locally owned or small chain supermarkets and less competitive discounters[ KMart, Shopko, etc.]

At 7/18/2008 3:39 PM, Anonymous Anonymous said...

matt_s said: "The emergence of a few Atlantic Superstores in Prince Edward Island, Canada ..., has been a negative for mom and pop grocers and general stores."

From the conclusion of the paper (did anyone actually read it?):

"There is no question that Wal-Mart does cause some mom-and-pop businesses to fail. However, those failures are entirely compensated for by the entry of other new small business elsewhere in the economy through the process of creative destruction."

Of course direct many competitors will fail. But they are replaced by new, better businesses, as shown in the paper, which is better for the economy as a whole (despite any sentimental connections the traditional general stores).

At 7/18/2008 3:42 PM, Anonymous Anonymous said...

Two linear regression lines for one data set? Ha.

At 7/18/2008 3:46 PM, Anonymous Anonymous said...

Thank you, sophist, for saying what I wanted to say. Old Perry has misused/misinterpreted statistics several times in the brief period I've been a reader. I'm glad someone has the smarts to take him to task over this sort of thing.

At 7/18/2008 4:06 PM, Anonymous Anonymous said...

"I know I'm a pain"

Wholeheartedly agree on that point there, sophist.

"Although Wal_Mart builds more stores in more populated areas, there are also more small business that survive frome the many more that would have survived had Wal_Mart not being there."

Not even Sir Humphrey Appleby could make sense of such gibberish. Is this a new form of English hitherto unknown to science?

How can one propose comparing small communities to larger population centres with the critical economic mass sufficient to support a Wallmart? Visited upstate NY recently?

"Statistics cannot be used as a weapon against common sense"

It seems you have found a plethora of such weaponry.

At 7/18/2008 4:15 PM, Blogger juandos said...

matt s says: "I will say that I speak from personal experience when I say that big stores do hurt small businesses, at least in rural areas"...

Are you sure small business people aren't hurting themselves?

Hmmm, a case of the blind leading the intensely blind? "Thank you, sophist, for saying what I wanted to say"...

"Statistics cannot be used as a weapon against common sense."...

Whereas common sense can't be bolstered by the misreading of statistics...

At 7/18/2008 5:38 PM, Blogger Jason Woertink said...

Does the study account for simply greater wealth? In other words a community that can support two Walmarts is probably wealthier than a community that supports only one so wouldn't such a community also support a larger number of small businesses?

At 7/18/2008 6:35 PM, Blogger Unknown said...

Anon wonders:

"Two linear regression lines for one data set? Ha."

Surprised? Not two, many. Least squares is one criterion to get the line.

Since the outliers in that scattered plot are skewed, the estimates are biased.

You can use contrained or unconstrained regularized least-squared methods, or weighted methods.

As far as the problem at hand, where did the linearity assumption come from? Non linear-least squares can produce multiple minimum solutions.

Many assumptions are required to be made to argue for linear least squares fit.

As far as I am concerned, the fit is a line with + slope to the middle and then a line with - slope to the other end.

Can you dictate how I weight my observations?


At 7/18/2008 6:39 PM, Blogger Unknown said...

"Whereas common sense can't be bolstered by the misreading of statistics..."

remember that

At 7/19/2008 6:51 AM, Anonymous Anonymous said...

A community with Wal-Mart has lower and middle income customers with increased disposable cash which they spend at other small businesses. Makes perfect sense.

The small grocer charging high prices for food goes out of business. Other small businesses prosper from the increased spending fueled by the savings.

At 7/19/2008 4:48 PM, Anonymous Anonymous said...

No linear regression model accounting for all of the data shown will produce a line of negative slope. Your wanting to apply a visual fit to override what the data portrays based upon what you call an outlier.

At 7/20/2008 10:14 AM, Blogger Unknown said...


You are talking about least squares linear regression. That, given the outlier to the left, will produce a positive slope.

But the choice of least squares is ad hoc and not dictated by laws of nature.

On the contrary, minimization of sum of Euclidean distances, which is essentially what least squares does for the totality of the points with respect to the line, is known to be in conflict with physical reality in many respects, as modern Topology has proven in the case of physical theories, such as electrodynamics, gravitomagnetics, and chromodynamics.

I mean, least squares is a too naive fit and introduces several biases, if not wrong alltogether.

At 7/21/2008 4:22 AM, Anonymous Anonymous said...

Statistics is a red herring.

Freedom is the issue, not math.

Wal-Mart should be free to conduct its business just as mom and pop proprietors are free to conduct their business.

One more thought. The anti-Wal-Mart crowd seem to be arguing that it is acceptable for a mom and pop business to charge high prices - but not the cable company and the oil company.

Exactly how small does a business have to be in order to price gouge? I want to keep my business just under this limit.

Is there a small business exception to windfall profits?

At 7/21/2008 10:25 AM, Blogger Marko said...

Sophist? More like Sophmore. A little philosphy is a dangerous thing. Yes, that is ad hominim. A+ in your Phil 101 class!

Sophist, you might want to look up the Straw Man informal fallacy in your little philosphy book. Also, Proof By Example, Hasty Generalization and others. You also make it sound like you think inductive reasoning is not sound. Bit of a problem for all of science, that.

The point remains that the data is the data, and doesn't seem to support the premise that walmart destroys the climate for small businesses.

If you own a small business that is threatened, you need to improve your business. Speciality butchers are thriving in many places, for example.

At 7/23/2008 10:54 AM, Blogger Christopher Monnier said...

As the study points out, the presences of a Wal-Mart store facilitates creative destruction. So direct competitors to Wal-Mart (such as small hardware stores, grocery stores, or low-to-middle-end clothing stores) will likely be put out of business by Wal-Mart, the money that Wal-Mart saves consumers will (as the study demonstrates) ultimately allow for the growth of other small business that don't directly compete with Wal-Mart. Moreover, residents of the community may, given the savings provided by Wal-Mart, have "extra" money to spend at the new small businesses.

At 7/23/2008 11:49 AM, Anonymous Anonymous said...

Wow, this study is really problematic.

Right off the bat, the article is only eight pages long. Perhaps in the upcoming publication in economic inquiry we'll get a better glimpse into the rigor of the authors' methodology, but I have never seen or even heard of a serious piece of research, respected in academia--especially an econ paper using quantitative data--which is this short and is mainly populated by graphs and conclusory statements.

There are also a number of value judgments built into this study. For example, the authors claim that benefits accrue to consumers in the form of money saved by shopping at wal-mart, where prices are lower than now defunct competitors. Unfortunately, behaviorial economics, which is THE ascendant field of study in economic academia today, shows that consumers are not self-interested rationally calculating automatons, but humans. As such, any benefit which accrues from lower prices is likely offset by the horrendous quality of the cheap good (i.e. lead toys from china), and behaviorial economics shows that this is not the result of rational cost-benefit analysis but rather human idiosyncracies. People will buy a 30 gallon jug of cyanide if they think it's a bargain.

It'd be nice if wal-mart really didn't have substantial negative effects on the communities it moves into, unfortunately this position paper isn't sufficient to persuade those who a) approach wal-mart with an open mind or b) approach wal-mart with hostility.
I look forward to the upcoming publication in economic inquiry the authors reference which will hopefully give us enough information to make an informed judgment on the validity of the study.

At 7/23/2008 12:01 PM, Anonymous Anonymous said...

The study is by a right wing "think" tank so its conclusions should immediately set a fair minded person to wondering about their motives. It does for me anyway. As one person noted, and I am no statistician, such a short paper with a few graphs must have limitations. I'll reserve judgement until the "full" version is published and I cant wait to see how more main stream and progressive folks evaluate the research.

At 7/23/2008 12:06 PM, Anonymous Anonymous said...

The last two comments (Christopher and Anon) really sum up this study well. Any serious social scientist will look at this study and question 1) why they didn't segment their measures of small and medium-sized businesses into those that compete directly with Wal-Mart (grocers, hardware, clothing, general stores) with those that do not (e.g., gas stations); and 2) why states rather than labor markets were used as the unit of analysis. Neither of these can be explained by a lack of data, but what appears to be lazy empirical design and a clear intention to find a specific result (you really think CATO would publish something to the contrary?). BTW- It doesn't take a behavioral economist (or what are more commonly called sociologists) to question this study--serious labor economists will likewise reject this study.

At 7/23/2008 1:37 PM, Blogger ryanconnercomedy said...

There's are two fundamental errors in this study.

1. Not examining population density. Population and population density are different, and have different implications.

Rural and urban trends are not the same. This accounts for the data appearing to be a buckshot.

2. The life of the small businesses are important data; more important than any instantaneous total of small businesses. We don't know if the businesses are thriving during Wal-Mart's existence, or if they're closing down and new businesses are moving in, only to close as well. That's the cycle in small towns that have been wrecked by Wal-Mart.

At 7/23/2008 6:07 PM, Blogger SBMB said...

The conventional wisdom as I know it isn't that Walmart is bad for small businesses in general, but that Walmart is bad for mom-and-pop retail stores. I grew up in a small town in Idaho. It used to have several drug stores, clothing stores, and hardware stores on the main street. Now the main street is mostly tourist shops and real estate offices. Are there fewer small businesses? Probably not. Are there fewer small retail businesses? Maybe not. But are there fewer retail businesses that sell necessities like food, clothing, medicine, hardware? Undoubtedly. Is this a net loss for the economy? I couldn't say. Is it a net loss for the culture? Undoubtedly.


Post a Comment

<< Home