European protected ham

ANOVA
linear regression
hierarchical model
In Europe, some types of ham—like Black Forest ham—can only be legally produced in specific geographic regions. Does the size of those regions affect the price of the ham?
Author

Alex Reinhart

Published

June 9, 2023

Data files
Data year

2021

Motivation

In the European Union, there are many foods that have long histories of being produced in one specific region. You may be familiar with champagne, produced in the historical province of Champagne in northeastern France; products made outside this region, even if they are very similar to champagne, cannot be called “champagne” and are usually called “sparkling wine”.

To protect these local foods, European Union regulations recognize certain “geographical indications” (GIs) for food, and restrict the names and labels that can be used for certain food items. There are two categories of GI:

  • A Protected Designation of Origin (PDO) product must be completely produced within a specific region. For example, only prosciutto from a specific region of Italy can be called “prosciutto di Parma”.
  • A Protected Geographic Indication (PGI) product must have at least one stage of production (but not necessarily all) within a specific region. For example, a product can only be called Black Forest ham if part of the production is done within the Black Forest region of Germany.

Additionally, a Traditional Speciality Guaranteed (TSG) product does not have to be from a specific region, but must be produced using certain traditional techniques or materials. For example, pizza napoletana (Neapolitan pizza) must be produced using a traditional method, but can be made anywhere.

Naturally, economists are interested in how this system affects the price of protected products. One theory is that a geographically indicated product’s price is inversely related to the area of the geographic region it can be produced in. If the area is very small, there can only be few producers, and the exclusivity of the product makes it seem more desirable or higher quality; if the area is large, there can be more producers, and the product seems less exclusive.

This dataset comes from a study meant to test this theory by using 22 types of GI ham (Höhn, Huysmans, and Crombez, 2023).

Data

The authors write:

We manually gathered data from online store websites operating in 11 EU countries, namely Austria, Belgium, France, Germany, Ireland, Italy, Luxembourg, the Netherlands, Portugal, Slovenia and Spain. These countries were chosen based on the following criteria. Our study encompasses all countries of origin of an eligible GI ham that have the euro as currency. Also, it includes the Netherlands as a major pig meat producer (Augère-Granier, 2020) and Ireland as a major ham importer (Török & Jambor, 2016).

We selected 36 online stores of established supermarkets that are present with physical stores in the respective city chosen for delivery, for example, Monoprix in Paris, REWE in Berlin or Coop in Rome. To avoid strong price differences due to strongly different store types we excluded specialty shops focusing on specific product categories. […] To ensure consistency we only gathered observations on supermarket websites providing home delivery in the respective capital’s centre. We collected these cross-sectional data in April 2021.

Each row of the data represents one type of ham from one particular online store. Of the 768 observations, 190 are GI hams (either PDO or PGI), while the rest have no geographic restrictions.

Data preview

european-ham.csv

Variable descriptions

Variable Description
COO Country of origin of geographical indication (GI) ham, or NA if ham is not GI
GI If ham has geographical indication (GI), 1, otherwise 0
GI_cluster Cluster at GI level, or NA if not GI. Cluster is the specific geographic indication, such as “Prosciutto di Parma”
GI_label Whether the ham is Non-GI, PDO (protected designation of origin), or PGI (protected geographical indication)
NonGI 1 if not labeled PDO or PGI, 0 otherwise
PDO 1 if the ham is PDO, 0 otherwise
PGI 1 if the ham is PGI, 0 otherwise
TSG 1 if the ham is TSG (traditional specialities guaranteed), 0 otherwise
breed 1 if a special pig breed is used for the ham, 0 otherwise
country Country in which the ham is sold
domestic 1 if the ham is domestic to the country, 0 if it is foreign
lnarea Natural logarithm of the GI area, square kilometers; NA for non-GI ham
lnproduction Natural logarithm of the total production of GI ham by the consortium, tons; only available for certain Italian GI hams, and NA otherwise
lnsourcing Natural logarithm of the area where the pig meat can be sourced, square kilometers
longevity Years passed since the PDO, PGI, or TSG designation was registered in the European Union, as of April 1, 2021
maturation The time the ham is matured, in months
national 1 if the ham is a national brand, 0 if it is a private label brand (i.e. made for the supermarket under their brand name)
organic 1 if the product is labeled/certified organic, 0 otherwise
packsize Size of the ham package, in grams
price Price of the ham per 100 grams, in Euros
producers Number of producers of GI ham in the consortium, for hams whose production is reported in lnproduction
smoked 1 if the ham is smoked, 0 otherwise
store Name of the online store the ham was sold by

Questions

  1. Are protected hams from different countries priced differently, on average? Conduct an ANOVA of price by country of origin (COO), using hams with a geographical indication. Interpret your results.
  2. The authors considered several covariates that also may influence prices: brand of the ham (a named national brand vs. supermarket brand), whether the ham is marked as organic, time taken for the ham to mature, and whether the ham comes from a special pig breed. Add these to your analysis. Comment on the results.
  3. To test their theory about geographic area, the researchers used the lnarea variable as part of their model. Fit a regression model predicting price using lnarea, the control variables from question 2, and the country of origin. Do the results support the theory?
  4. One concern may be that ham prices within a country of origin may be dependent. One could produce a hierarchical model, where each ham has a mean price that depends on the country of origin and also the geographic area of production. Build a suitable hierarchical model and justify your model choices. Fit the model and interpret your results.

References

Gero Laurenz Höhn, Martijn Huysmans & Christophe Crombez (2023). “Does terroir size matter? Protected geographical areas and prices of European hams.” Regional Studies. https://doi.org/10.1080/00343404.2023.2187365

Gero Laurenz Höhn, Martijn Huysmans & Christophe Crombez (2023). “Replication Data for: Does Terroir Size Matter? Protected Geographical Areas and Prices of European Hams”, https://doi.org/10.34894/RX0QIN, DataverseNL, V1.