Consumers' willingness to pay for state animal welfare labeling: Evidence from a stated choice experiment

Last registered on August 26, 2021

Pre-Trial

Trial Information

General Information

Title
Consumers' willingness to pay for state animal welfare labeling: Evidence from a stated choice experiment
RCT ID
AEARCTR-0008079
Initial registration date
August 23, 2021
Last updated
August 26, 2021, 9:35 AM EDT

Locations

Region

Primary Investigator

Affiliation
University of Hamburg

Other Primary Investigator(s)

Additional Trial Information

Status
In development
Start date
2021-08-27
End date
2021-09-09
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
Animal farming for the purpose of meat and dairy production, especially in the most prevalent intensive way, has significant environmental, health and ethical consequences, which are increasingly debated in politics and public. Apart from the fundamental question of eating animals, an increasing number of meat eaters questions under which conditions farming animals are reared.
Survey results show that farm animal welfare is important to a majority of consumers (BMEL 2021). These preferences are, however, not reflected in the market yet. From an economics perspective, animal welfare is a public good (Grethe 2017), leading to potential free-riding behavior and negative externalities for those agents who care about animals’ well-being (Harvey and Hubbard 2013; Lusk and Norwood 2011). Moreover, animal welfare is a credence attribute for consumers (Darby and Karni 1973) that cannot be verified after purchase or even after consumption. The resulting information asymmetry contributes to a failure of the market for animal welfare (Lusk et al. 2007). Even those consumers who are interested in contributing to the public good cannot do so because they lack information on the animal welfare conditions underlying the products. Labeling is an option to provide (additional) information to consumers (Teisl and Roe 1998), transferring improvements towards the social optimum to consumers. Improvements need to be accompanied by a higher willingness to pay (WTP) for better animal welfare conditions.
The German government plans to introduce a state animal welfare label ("Tierwohlkennzeichen") for fresh meat products for which farmers can voluntarily apply if they rear their farm animals under better animal welfare conditions than the ones defined by legal minimum standard (BMEL 2020). The label is supposed to consist of three levels and 13 criteria that have to be fulfilled for each level are already defined. Three research questions arise from this political and public discourse.

Research on labeling certain aspects of husbandry conditions and on fictitious animal welfare labels finds a positive WTP for labeled meat products in choice and contingent valuation experiments (e.g., Carlsson et al. 2020; Yang and Renwick 2019; Clark et al. 2017; Gracia and de-Magistris 2016; Janssen et al. 2016; van Loo et al. 2014; Koistinen et al. 2013). Most of the studies, however, test a fictional label or husbandry criteria that researchers picked themselves as opposed to an actually planned state label with its specifics. The first research question is thus: Are consumers willing to pay for a German state animal welfare label with all its complexities compared to no label?

In addition, only fresh meat is planned to be labeled, although processed meat products, such as sausages or bacon, account for half of the average meat consumption per person in Germany (DFV 2020). Hence, the second research question is: Does the WTP for the label differ depending on product type, i. e. whether it is provided for fresh or processed meat?

On top of the labeling, policy makers discuss how to finance the transformation in the farming sector towards higher animal welfare standards. The most favored option so far has been to introduce an additional consumption tax for meat products, a so called animal welfare tax (“Tierwohlabgabe”; Kompetenznetzwerk Nutztierhaltung 2020). The tax is supposed to be collected on every kilogram of meat sold (0.40 EUR/kg) independent from the type of meat and animal welfare conditions under which the meat has been produced. Earnings from the tax are planned to be redistributed to farmers who implement higher animal welfare standards. As a result, every meat buyer would contribute at least a little towards animal welfare and every consumer knows that others are contributing, too. The potential free-riding behavior resulting from animal welfare’s public good character might thereby be influenced. To our knowledge, the interaction between an informational and a tax instrument in the context of animal welfare has not been researched yet. These considerations lead to the third research question: Does the introduction of an additional animal welfare consumption tax change the WTP for the state animal welfare label?
External Link(s)

Registration Citation

Citation
Schwickert, Henrike. 2021. "Consumers' willingness to pay for state animal welfare labeling: Evidence from a stated choice experiment." AEA RCT Registry. August 26. https://doi.org/10.1257/rct.8079-2.0
Experimental Details

Interventions

Intervention(s)
The experiment comprises a between- and a within-subjects design.
There are two experimental groups for the between-subjects design: One group has to choose among fresh pork meat products ("Schnitzel" treatment), the other one among processed pork meat products (sausages, "Würstchen" treatment).

Within the two experimental groups, respondents are primed with two different scenarios in randomized order: One block of choice sets in which no additional animal welfare tax is imposed on the products and one block in which an additional animal welfare tax of 0.40 EUR/kg is imposed on each product offered.
Intervention Start Date
2021-08-27
Intervention End Date
2021-09-09

Primary Outcomes

Primary Outcomes (end points)
Discrete choice experiment among three alternatives and a no-choice option. Each alternative has three attributes. The attributes are origin of meat (qualitative, 2 levels: regional/Germany), price per 350g of meat (quantitative, 2.79 EUR to 10.99 EUR), labeling of the product (qualitative, 6 levels: no label, state animal welfare label 1 star, 2 stars, 3 stars, 3 stars + German state organic label, German state organic label only).

The key outcome variable is a subject's choice between the three alternatives and the no-choice option in a choice set.
Variable "choice" is equal to 1 for the chosen option in a choice set and 0 for the options not chosen. Using discrete choice models, namely simulated estimation of mixed logit models, the aim is to derive the estimated willingness to pay (WTP) for the state animal welfare label and its different levels. The corresponding price attribute is the price per 350g of meat.
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
The stated choice experiment is part of an online survey.
In a first step, respondents representative for the German population in terms of age, gender and region of living (on a federal state level) are screened for two criteria to have access to the survey: 1. being at least partially responsible for grocery shopping in own household; 2. buyers of pork meat. If respondents fulfill the screening criteria, they are randomly allocated to one of the two experimental groups (fresh meat "Schnitzel" treatment or processed meat "Würstchen" treatment) and to one of the two scenarios to see first ("No tax" scenario or "Tax" scenario). Thus, there are four combinations to which they can be assigned: 1. "Schnitzel" treatment + "No tax" scenario first; 2. "Schnitzel" treatment + "Tax" scenario first; 3. "Würstchen" treatment + "No tax" scenario first; 4. "Würstchen" treatment + "Tax" scenario first.

The stated choice experiment starts with an introduction of the (hypothetical) choice setting, namely that respondents should imagine to go grocery shopping for meat products (either 350g of fresh meat "Schnitzel" or 350g of processed meat "Würstchen") in a supermarket. Before they see the first choice set, respondents are provided with information on the three attributes. If they are in the "Tax" scenario, they are informed that all prices include an animal welfare consumption tax of 0.40 EUR per kg, i. e. 0.14 EUR per 350g.

Respondents receive twelve choice sets in total, grouped into two blocks, which are shown in random order: one block with six choice sets in a scenario without a tax and another block with six choice sets in which the price contains the animal welfare consumption tax.

After the stated choice experiment, debriefing questions on perception of consequentiality, attitudes towards animal farming and own moral obligations, knowledgeability about and perception of the state animal welfare label as well as socio-demographics follow.


The design is supposed to test the following hypotheses:
To measure consumers’ WTP for labels and for their underlying attributes, choice experiments are common methods. Labels have shown to increase WTP for certain (public good) attributes of food products, for example for certain husbandry conditions, carbon footprint, organic, health and nutritional facts, quality grades or a combination of those (e.g., Macdiarmid et al. 2021; Carlsson et al. 2020; Lusk et al. 2018; Gracia and de-Magistris 2016; Van Loo et al. 2014; Koistinen et al. 2013). Research also finds that the state is the preferred and credible institution to provide labels in the food context as opposed to retailers, industry organizations, NGOs and consumer groups (Gracia et al. 2011; Olynk et al. 2010; Vanhonacker et al. 2010). We thus hypothesize:

H1a: Respondents have a higher WTP for any level of the state animal welfare label compared to no label.

With regard to the three label levels, previous research on multi-level labels finds decreasing marginal WTP with increasing animal welfare levels, indicating demand for medium animal welfare conditions as opposed to only premium levels (Weinrich and Spiller 2016; De Jonge et al. 2015; Heerwagen et al. 2015). We hence expect similar findings, amending the first hypothesis to:

H1b: The increase in respondents' WTP is marginally decreasing from the lowest to the highest animal welfare level.

If the objective is to resolve the information asymmetry, labels should be used on all products impacting the public good of animal welfare, independent from their degree of processing. A rational agent should have the same WTP for animal welfare independent from the labeled product if it originates from the same animal. However, psychological research provides evidence that the processing level of the meat product makes a difference: The larger the dissociation of the meat product from the animal it stems from, the lower the empathy for the animal (Kunst and Hohle 2016; Hoogland et al. 2005). Following this logic, the dissociation from the animal should be larger in the case of sausages, a highly processed meat product, compared to fresh meat. We therefore divide participants into two treatment groups, the fresh meat "Schnitzel" treatment and the processed meat "Würstchen" treatment and test the following hypothesis:

H2: Respondents' WTP for any label level is higher in the "Schnitzel" treatment than in the "Würstchen" treatment.

While labeling could (partly) resolve information asymmetry, it does not solve the problem of free-riding (Lusk et al. 2007). Three previous studies explicitly question if such an externality in the case of animal welfare actually exist. Carlsson et al. (2007) and Tonsor et al. (2009) do not find a free-rider deficit in the case of battery cages in egg production and gestation crates for pigs respectively, while Uehleke and Hüttel (2019) detect free-riding behavior in their contingent valuation study. The political plans to introduce a consumption tax for animal welfare also raises the question if this introduction impacts or reveals free-riding behavior. As the tax is supposed to be imposed on every kilogram of meat sold, independent from its type and the animal welfare conditions under which it has been produced, it has to be paid by every consumer who decides to buy a meat product. Two reactions are possible, which we test confronting each participant with two scenarios, one without a tax and one with the animal welfare tax included in the price.
On the one hand, every individual now knows that others have to pay a tax, too. As the tax is earmarked to be paid out only to farmers who provide higher animal welfare standards, consumers automatically contribute to the public good of animal welfare when buying a meat product. This could lead to consumers being motivated to additionally contribute to animal welfare by buying labeled products and free-riding might be reduced. Their WTP for the animal welfare label would increase. The result would be similar to the effect found in Uehleke and Hüttel (2019) for the voting on product bans or Mulder and Zomer (2017) for providing information to respondents on market shares of animal welfare products.
On the other hand, people could perceive the payment of the consumption tax as some kind of indulgence. By paying the tax, they already contribute to the public good of animal welfare. Thus, there is no need to conduce further by buying products with labeled animal welfare. In this case, the tax would encourage free-riding behavior. Consumers’ WTP for the animal welfare label would decrease compared to the “No tax” scenario. Both directions are conceivable. We thus hypothesize:

H3: Respondents' WTP for any label level is different between the "Tax" treatment and the "No tax" treatment.
Experimental Design Details
A few details on the design of the stated choice experiment:
Three measures are taken to address potential hypothetical bias ex ante. Using a cheap talk script (Murphy et al. 2005; List 2001; Cummings and Taylor 1999), respondents are reminded to consider their budget when choosing among alternatives. In addition, before seeing the first choice set, respondents are asked if they can, "hand on heart" (common idiom in German, meaning "to cross one's heart"), promise to decide as if they were in a real shopping situation (Carlsson et al. 2020; Jacquemet et al. 2013). Finally, providing a no-choice option makes the situation more realistic as in a real supermarket respondents would not be forced to choose between offered alternatives (Adamowicz et al. 2011).

In the introduction of the choice task, respondents have the option to receive more detailed information on the underlying criteria for the state animal welfare and organic label by clicking on a small image, which is then enlarged. In the background, the software measures if respondents clicked on the image and for how long it is enlarged.
Randomization Method
Randomization will be done by Sawtooth Software used for programming survey and experiment. Each respondent is allocated to one of the four combinations (as described in "Experimental Design" section) based on a random draw from numbers 1 to 4.
Randomization Unit
Unit of randomization will be individual respondents.
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
1,200 individuals, design is not clustered
Sample size: planned number of observations
1,200 individuals, living in Germany
Sample size (or number of clusters) by treatment arms
Fresh meat ("Schnitzel") treatment: 600
Processed meat (sausages, "Würstchen") treatment: 600
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Supporting Documents and Materials

There are documents in this trial unavailable to the public. Use the button below to request access to this information.

Request Information
IRB

Institutional Review Boards (IRBs)

IRB Name
Dean's Office of the Faculty of Economics and Social Sciences at University of Hamburg
IRB Approval Date
2021-08-19
IRB Approval Number
N/A, copy of the document attached

Post-Trial

Post Trial Information

Study Withdrawal

There are documents in this trial unavailable to the public. Use the button below to request access to this information.

Request Information

Intervention

Is the intervention completed?
No
Data Collection Complete
Data Publication

Data Publication

Is public data available?
No

Program Files

Program Files
Reports, Papers & Other Materials

Relevant Paper(s)

Reports & Other Materials