Rafi Mohammed

How Much Does Putting Those Little Price Tags on Grocery Items Really Cost?

Posted on March 20th, 2007 (0 Comments)

On March 10, Paul Rubin wrote an editorial in the Wall Street Journal titled “A Penny Not Saved” that summarized a research paper he co-authored (along with Mark Bergen, Daniel Levy, Sourav Ray, and Benjamin Zeliger) titled “When Little Things Mean a Lot: The Inefficiency of Item Pricing Laws.” Rubin’s editorial highlights the costs of consumer regulation. In the United States, some states have item pricing laws (IPL) which require most items in a store to have its own price sticker. Other states that don’t have IPLs (NO IPL) simply require a price tag on the shelf. A key reason for having IPLs is because price tags help consumers check if pricing errors are being made at the cash register.

Studying grocery store prices in states with and without IPLs, Rubin and his co-authors came to an astonishing conclusion: prices in states with IPLs were on average 20 – 25 cents higher than states without IPLs. Given that the average product price in their survey was $2.50, that’s roughly a 10% premium. The authors fully attribute this premium to the consumer protection regulation of IPLs. After explaining this study to my friend George, he jumped out of his chair and screamed “WOW.” I’m sure that is what many of you are thinking, 20 – 25 cents per item is a lot of money.

This blog is not really about my opinion on IPLs (I’m agnostic), but more about questioning the claims that experts make to you. In my line of work, generating new price strategies for non- and for-profit organizations, I encourage questions. If I cannot answer a question or explain my work in a manner that clients understand, I’ve failed. No question is stupid – it’s my job to clearly explain my methodologies…I believe all experts should be held to this standard.

After thinking about Rubin et al’s results, I started wondering why it costs 20 – 25 cents more per item to have someone with a price stamping “gun” paste labels onto individual products. You’ve probably seen those guys at supermarkets - they can really get into the groove and quickly stamp out those labels. While I don’t doubt the 20 – 25 cent price differential in their survey, I question whether it’s fair to solely attribute this price differential to IPLs. Returning to my favorite pricing framework, The Value Decoder, here are some basic questions that I have about Rubin et al’s analysis.

  1. What about income? The authors claim they studied stores in similar markets (socio-economically, same/similar supermarkets in size, type etc). But in one of their surveys, they compared IPL prices in Tarrytown, NY to No-IPL prices in Clifton, NJ. A quick check with 2000 US Census data found the median income in Tarrytown was $68,762 and $50,619 in Clifton City, New Jersey. Perhaps Tarrytown’s 36% higher median income explains why prices are higher in Tarrytown, NY (an IPL state) relative to Clifton, NJ (a No-IPL state).
  1. What about competition? To the best of my knowledge, the authors do not account for the role of competition in their study. Prices are probably lower in areas where two strong rivals are located close to each other compared to less competitive areas. Not controlling for competition leaves open the possibility that high-priced stores in Rubin et al’s fairly small sample (4 stores in their first survey, 16 new stores in their second survey) could simply be located in areas with low competition.

Just to be clear, my goal is to not discredit the Rubin et al. paper. I should note their paper will be published in the Journal of Law and Economics, a prestigious academic journal. Research is a tough gig – it’s rare to produce a paper where questions like mine aren’t asked. In an in-person seminar, you ask such questions, listen to the author’s response and then decide whether or not to believe their explanation. You have to admit, my questions are pretty basic (probably ones that many of you would think “stupid” to ask in a public setting) – and I think the authors should have addressed them in their paper.

I did email the authors last week to address these questions. In their reply, they stated that they’d try to get back to me later in the week. But I haven’t heard back from them. I’d love to hear back from them and publish their response – this is not a question of me or they being wrong – it’s about clarifying some points on very important research.

You can download Rubin et al’s paper at SSRN.com by searching under Paul Rubin’s name and the title “When Little Things Mean a Lot: On the Inefficiency of Item Pricing Laws.”

Please feel free to send me any questions or comments. I update this blog two to three times a week – please consider signing up to be notified by e-mail of a new blog post. Thanks for taking the time to read my blog.

Add Comment
Send to Friend
Email Signup
RSS Feed