Skip to main content

Advertisement

Log in

What is an Unregulated and Potentially Misleading Label Worth? The case of “Natural”-Labelled Groceries

  • Published:
Environmental and Resource Economics Aims and scope Submit manuscript

Abstract

Motivated by a multitude of lawsuits and considerable public policy debate in the U.S., this study provides insight on the demand effects for an unregulated and potentially misleading phrase found on many grocery labels: “all natural”, and its common variations. Using a targeted sample of adult grocery shoppers, we employ an incentive-compatible approach to elicit the willingness to pay for a variety of “natural”-labelled food products, along with their counterfactual, standard-labelled counterparts. We find that consumers are willing to pay 20% more on average for “natural” products. Using elicited information on consumer beliefs, we find that this premium decreases when “natural” signals no artificial flavors or preservatives, and increases when consumers believe that “natural” means GMO-free. Interestingly, for those indicating that the “natural” designation is meaningless, they are willing to pay about one-third less for products labelled this way.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3

Similar content being viewed by others

Notes

  1. The phrase “natural” is regulated for a subset of foods. In particular, the USDA defines “natural” for meat, poultry and eggs as products that are “minimally processed” and contain no artificial ingredients.

  2. The FDA has occasionally sent warning letters to manufacturers who used a clearly synthetic ingredient in their “natural” product. However, these letters often go ignored, and “deceptive” products have been allowed to remain in the marketplace (Silverglade and Heller 2010).

  3. Meanwhile, the number of “natural” label class-action lawsuits continues to rise. However, some courts have delayed ruling in hopes of an FDA decision (Menayang 2016). The delay has prompted several companies to settle, avoiding lengthy litigation, such as Good Karma Foods. Other lawsuits have been dismissed by judges, but there are still many cases waiting “in the pipeline” (Allen et al. 2013; Frankel 2013, and Smith 2014).

  4. In particular, the law allows companies to use QR codes, toll-free phone numbers or URLs rather than a transparent, on-package label. It has yet to be seen how courts will factor this into the many lawsuits addressing both “natural” and GMOs (Watson 2016). See also Cummins (2014).

  5. The dominant strategy to bid one’s maximum WTP is not altered by the underlying price distribution, as long as this value is within the (perceived) range of possible prices.

  6. Cason and Plott (2014) provide evidence that non-optimal choices in an induced-value BDM experiment are related to the (posted) maximum of the random price range. The use of an uncertain maximum price, along with other experimental procedures discussed below, serve to avoid the misconceptions identified in their study.

  7. In particular, induced value experiments typically describe the participant as having a “redemption value” for a fictional good, such as a “ticket” or a “widget” (see, for example, Irwin et al. 1998). In a pilot session we found that such constructs were confusing to some participants. As discussed by Irwin et al. (1998), the feedback provided by playing out a BDM procedure is not strong; specifically, there is usually a low chance a participant is punished from placing a bid that deviates even moderately from induced value. With this concern in mind, participants place a bid in a training round, and then the purchase procedure is played out four times to allow for both learning and additional earning opportunities. The random price is further drawn from a fairly tight distribution, a four-dollar range centered on induced value. The price distribution is not announced prior to bidding, although the instructions emphasize that losing money and gaining money are both possible outcomes.

  8. For instance, the majority bid exactly $3.00 (the optimal bid) or within 10% of $3.00 in the training round where $3.00 is auctioned. As further (suggestive) evidence that participants understood the instructions, more than 90% of participants responded “4” or “5” to the question, “Did you understand the instructions for the experiment today? Please rate your understanding on a scale from 1 to 5 where ‘5’ was ‘I understood very well”’.

  9. At present no GM wheat is commercially grown in the US for consumption. However, wheat crackers contain many ingredients other than wheat (e.g., canola oil and corn syrup), which are almost always made with GMOs.

  10. The use of actual labels would have introduced additional variation in observed and unobserved product attributes (e.g. brand familiarity). However, the researcher-generated labels are completely accurate and reflect information on the actual product labels. The researcher-generated labels for “natural” products state either “natural”, “all-natural”, or “100% natural”, depending on what appears on the actual product label. As noted by a reviewer, the actual phrase used may be an important driver of “natural” premiums. This remains an open research question.

  11. We were able to find some of the product pairs in the exact same size containers. However, if the containers were similar, but not exactly the same, we included approximate sizes on the researcher-generated label. We always used an approximate size that was smaller than or equal to the size of the actual product. This way, participants would never pay more than they intended for a product, given its size.

  12. The actual product brands used were included on the labels provided to the participants, with the exception of the “natural” toothpaste (Tom’s). For this case, we felt the brand name would draw consumers’ attention since Tom’s products are popular and are always advertised as “natural”. It is for this reason we appended the phrase “or similar” to the list of brands on all the labels.

  13. The recent demographic data we report was not available at the time of our recruitment efforts. Our targeting efforts based on sex were driven by data from 2001, which suggested that a much higher percentage of shoppers are women.

  14. Analysis was conducted with Stata version 14.

  15. It is more common to analyze auction bids with a Tobit model, treating $0 bids as corner solutions. In this study we instead use linear regressions, as just 8% of observations correspond with $0 bids and OLS is more robust to model misspecification than Tobit. Moreover, assuming that bids are a linear function of covariates is a reasonable approximation, as covariates are mainly indicator variables. For the interested reader, in the online supplementary material we present Tobit models that parallel the OLS regressions in Table 4.

  16. Specifically, we estimated separate probit models for each belief designation, including the first nine variables listed in Table 2 as covariates.

  17. Part 2 bids may be partially driven by income effects, given that participants in Part 2 are bidding with the knowledge that one of the Part 1 decisions is binding. These effects should be negligible, however, as based on Part 1 bids there is roughly a one-in-three chance of making a purchase, the odds of buying a product in both parts is less than ten percent, and earnings from the experiment are high relative to purchase prices. On average, Part 2 bids are ten cents lower than for Part 1.

References

Download references

Acknowledgements

We thank the co-editor and two anonymous referees for providing thorough and constructive feedback on earlier versions of this paper.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Christian A. Vossler.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary material 1 (docx 335 KB)

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Butler, J.M., Vossler, C.A. What is an Unregulated and Potentially Misleading Label Worth? The case of “Natural”-Labelled Groceries. Environ Resource Econ 70, 545–564 (2018). https://doi.org/10.1007/s10640-017-0132-9

Download citation

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10640-017-0132-9

Keywords

JEL Classification

Navigation