Mental Accounting, Information Avoidance
and Voluntary Carbon Offsets

Nicola Campigotto
nicola.campigotto@unitn.it
Chiara Gioia
chiara.gioia@unitn.it
Matteo Ploner
matteo.ploner@unitn.it

June 2024

Carbon what?

Ryanair

Flixbus

Shopify

Carbon what? (cont’d)

  • Voluntary carbon offsets (COs) are tradable certificates linked to climate projects.

  • By purchasing COs, individuals and firms can fund the projects (reforestation, energy efficiency, renewable energy, …) associated with them.

  • Huge market: in 2021, the voluntary carbon market reached $2 billion — four times its value in 2020. The pace of purchases continued to accelerate in 2022.

This paper

  • The literature has mainly focused on estimating individuals’ willingness to pay for COs (e.g. Berger et al. 2022; Rodemeier 2023).

  • The psychological factors influencing carbon offsetting behaviour are still understudied.

  • We focus on two of these factors

    • Mental accounting (MA): the cognitive bookkeeping process by which individuals evaluate money transactions.
    • Strategic ignorance (SI): people’s deliberate avoidance of information that might cause them distress.

Overview

  • Two pre-registered (osf.io/9y3dj), incentivized online experiments (named MA and SI) conducted on Prolific.

  • Sample: 1000 UK residents.

    • 80% power with a 5% significance level to detect an effect size of about 0.25 using a two-sided Wilcoxon-Mann-Whitney test 💪

  • No deception.

  • 1.5 GBP participation payment (~13.8 GBP/hour) + bonus payment (Ryanair/Amazon voucher worth up to 100 GBP).

Overview (cont’d)

  • Each experiment consisted of three phases:
    1. Random assignment to treatment conditions + survey.
    2. Choice whether to contribute (and, if so, how much) to a carbon offset programme.
    3. Elicitation of attitudes towards two articles discussing the environmental impact of flying and e-commerce.

Offsetting decisions

  • CO contributions were made on a voluntary basis.

  • Subjects were asked whether they would like to participate in a CO programme.

  • Those willing to participate had to specify their desired contribution within a range of 0.1 to 10 GBP, with intervals of 0.1 GBP.

  • Participants were informed that contributions would be deducted from the bonus prize.

    • An individual who contributed \(x\) and was then selected to win the prize would receive a voucher worth \(100 - x\) GBP.

Offsetting decisions (cont’d)

  • Contributions from subjects selected to win the bonus prize were transferred to Trees for All, a non-profit organization that restores forests around the world.

  • Contributions from subjects not selected to win the bonus prize were not transferred.

  • Thus, actual contributions were contingent on winning the prize (incentive compatibility requirement).

Treatments: Study MA

  • The reatments only differed in the bonus prize.

  • Treatment conditions

    1. Amazon_noinfo (n = 200). Bonus prize: Amazon voucher.
    2. Ryanair_noinfo (n = 200): Bonus prize: Ryanair voucher.
  • No information about the emissions associated with the bonus prize.

Treatments: Study SI

  • The treatments differed in the information given to subjects when deciding whether to contribute to the CO programme.

  • Treatment conditions

    1. Ryanair_info (n = 200). Subjects were given an estimate of the per-flight and per-passenger emissions associated with a 100 GBP Ryanair ticket.

    1. Ryanair_noinfo (n = 200). Subjects were not given any such information. Shared with Study MA.

    2. Ryanair_choose (n = 400). Subjects could choose whether or not to be informed.

Treatments: Study SI (cont’d)

  • Subjects in Ryanair_choose sorted themselves into two sub-conditions, Ryanair_choose_info and Ryanair_choose_noinfo.

  • Those who chose to receive information had to complete a real-effort decoding task.

  • Subjects could quit the decoding task at any time, in which case they received no information.

Treatments: overview


Main hypotheses

  • Question: are contributions to the CO programme more likely to be allocated to the same mental account as the Ryanair voucher than the Amazon voucher?
    • Conjecture: flying is more readily associated to carbon offsetting than online shopping.
    • Due to this cognitive proximity, the gain from the voucher and the offsetting cost are more likely to be allocated to the same mental account in the Ryanair_noinfo condition.

Hypothesis 1 (Study MA)

Subjects in Ryanair_noinfo contribute more to the CO programme than subjects in Amazon_noinfo.

Main hypotheses (cont’d)

  • Question: does increasing the salience of the environmental impact of people’s actions have a positive effect on contributions?
    • Conjecture: receiving information about the impact of flights raises sustainability concerns.

Hypothesis 2 (Study SI)

Subjects in Ryanair_info contribute more to the CO programme than subjects in Ryanair_noinfo.

Main hypotheses (cont’d)

  • Question: does the decision to receive information about the impact of flights affect offsetting behaviour?
    • Conjecture: individuals with a low or moderate propensity to contribute choose not to receive information and donate little.
    • These individuals would feel compelled to donate more if they had the information.
    • Thus, they strategically choose to ignore the information in order to donate less and avoid cognitive dissonance (Bénabou and Tirole 2016).

Main hypotheses (cont’d)


Hypothesis 3 (Study SI)

Individual characteristics that predict lower contributions to the CO programme are associated with a higher likelihood of self-selecting into Ryanair_choose_noinfo.

Hypothesis 4 (Study SI)

Contributions in Ryanair_choose_info are higher than contributions in Ryanair_info, as individuals with strong environmental concerns tend to self-select into choose_info.

Contributions in Ryanair_choose_noinfo are lower than contributions in Ryanair_noinfo, as individuals with little environmental concerns tend to self-select into choose_noinfo.

Results

Result 1 ✅

Subjects in Ryanair_noinfo contribute significantly more to the CO programme than subjects in Amazon_noinfo (two-sided Wilcoxon test p-value < 0.001).

Determinants of contributions, Study MA
(1)
OLS		 (2)
Tobit
Constant 3.813
(3.397)		 -2.831
(14.547)
Treatment: Ryanair_noinfo 1.545***
(0.401)		 6.619***
(1.752)
Risk tolerance −0.094
(0.103)		 −0.521
(0.103)
Patience −0.033
(0.122)		 0.021
(0.517)
Social concerns 0.421***
(0.107)		 1.813***
(0.470)
Environmental attitudes 0.435***
(0.116)		 1.845***
(0.504)
Environmental concerns 1.418**
(0.473)		 5.486**
(1.995)
Male 0.841
(0.438)		 3.496
(1.837)
Age −0.020
(0.020)		 −0.055
(0.083)
Additional controls yes yes
Adj. R2 0.198
Observations 399 399§


\(^{***}\)p < 0.001; \(^{**}\)p < 0.01; \(^{*}\)p < 0.05

§: 105 left-censored observations, 203 right-censored observations.

Dependent variable: contribution to the CO programme.

Additional controls: Ethnicity, Education, Subjective socioeconomic status, Employment or student status, Number of approvals on Prolific.

Results (cont’d)

Result 2 ❌

Subjects in Ryanair_info do not contribute significantly more to the CO programme than subjects in Ryanair_noinfo (two-sided Wilcoxon test p-value = 0.089).

Results (cont’d)

Result 3 ✅

Subjects who are more future-oriented and concerned about climate change are more likely to sort themselves into Ryanair_choose_info.

Determinants of self-selection into Ryanair_choose_info, Study SI
Logit
Constant -1.593
(1.797)
Environmental attitudes 0.211**
(0.072)
Risk tolerance -0.031
(0.054)
Patience 0.181*
(0.071)
Social concerns 0.016
(0.071)
Environmental concerns 0.349
(0.263)
Male 0.067
(0.240)
Age -0.017
(0.011)
Additional controls yes
Observations 400


\(^{***}\)p < 0.001; \(^{**}\)p < 0.01; \(^{*}\)p < 0.05

Additional controls: Ethnicity, Education, Subjective socioeconomic status, Employment or student status, Number of approvals on Prolific.

Results (cont’d)

Result 4 ✅

Contributions in Ryanair_choose_info are significantly higher than in Ryanair_info (two-sided Wilcoxon test p-value < 0.001).

Contributions in Ryanair_choose_noinfo are significantly lower than in Ryanair_noinfo (two-sided Wilcoxon test p-value < 0.001).

Determinants of contributions, Study SI
(1)
OLS		 (2)
Tobit
Constant -1.046
(1.993)		 -28.940*
(11.890)
Treatment: Ryanair_noinfo 0.499
(0.401)		 3.587
(2.333)
Treatment: Ryanair_choose_info 0.840*
(0.426)		 5.708*
(2.538)
Treatment: Ryanair_choose_noinfo -0.752
(0.392)		 -3.934
(2.249)
Risk tolerance -0.158*
(0.069)		 -0.994*
(0.407)
Patience 0.128
(0.085)		 0.829
(0.498)
Social concerns 0.489***
(0.079)		 2.892***
(0.508)
Environmental attitudes 0.498***
(0.078)		 2.953***
(0.519)
Environmental concerns 0.751*
(0.325)		 3.620
(1.887)
Male 0.159
(0.304)		 1.305
(1.777)
Age 0.033*
(0.014)		 0.184*
(0.082)
Additional controls yes yes
Adj. R2 0.197
Observations 799 799§
Test: Ryanair_noinfo – Ryanair_choose_noinfo = 0 (p-value) <0.001 <0.001


\(^{***}\)p < 0.001; \(^{**}\)p < 0.01; \(^{*}\)p < 0.05

§: 212 left-censored observations, 455 right-censored observations.

Dependent variable: contribution to the CO programme.

Additional controls: Ethnicity, Education, Subjective socioeconomic status, Employment or student status, Number of approvals on Prolific.

Wrap-up

  • Monetary transactions related to activities that are clearly perceived as polluting are more likely to be allocated to the same mental account as carbon offsets, resulting in higher offset contributions.

  • Individuals with a low pro-environmental orientation tend to strategically ignore information about the emissions associated with their activities, so as to avoid cognitive dissonance and contribute less.

  • The choice to actively avoid (seek) information undermines (reinforces) pro-environmental behaviour relative to situations in which information is unavailable (compulsively provided).

References

Bénabou, Roland, and Jean Tirole. 2016. “Mindful Economics: The Production, Consumption, and Value of Beliefs.” Journal of Economic Perspectives 30 (3): 141–64.
Berger, Sebastian, Andreas Kilchenmann, Oliver Lenz, and Francisco Schlöder. 2022. “Willingness-to-Pay for Carbon Dioxide Offsets: Field Evidence on Revealed Preferences in the Aviation Industry.” Global Environmental Change 73: 102470.
Rodemeier, Matthias. 2023. “Willingness to Pay for Carbon Mitigation: Field Evidence from the Market for Carbon Offsets.” IZA Discussion Paper 15939.