Molecular and Cellular Mechanism of Action of Hormones such as Growth Hormone...
Plastic bags and prospect theory; Lessons in policy design
1. Plastic bags and prospect theory;
Lessons in policy design.
Peter King
PhD student, Department of Economics, University of Bath
Paper: bit.ly/Latte-Levy
2. Motivation
• Half a million coffee cups littered every day
• Producers want a discount scheme
• Government wants a tax scheme
• Many parallels to the `plastic-bag tax’
• Summary: Best solution to regulate externality-causing goods.
3. Research questions
• Which instrument is more effective?
• Prediction: Tax more effective due to framing effects
• How strong are behavioural spillovers?
• Prediction: Plastic bag tax improved attitudes towards future policies.
• Can response bias be mitigated in a survey of hypothetical
willingness-to-pay estimates?
• Prediction: Survey estimates are inflated upwards due to bias.
4. Literature
• Discount schemes
• New to the literature, only limited evidence available at all.
• However, substantial evidence for subsidies and other similar incentives
• Pigouvian taxation
• Well supported theory and design
• Empirical support from South Africa, Ireland, and Wales where similar taxes
were implemented.
• Convery, McDonell and Ferreira (2007), Dikgang (2010), Poortinga,
Whitmarsh and Suffolk (2013)
5. Prospect theory
• Theory:
• Individuals are risk-averse in the domain of perceived gains and risk-
seeking for perceived losses to the status quo
• Individuals will do more to avoid a loss than to achieve a gain
• Frame the tax as a loss and the discount as a gain
• Prediction that agents react more strongly to tax than discount
• Evidence
• Tversky and Kahneman (1981)
• Poortinga, Whitmarsh and Suffolk (2013)
• Relevance
• Framing of policy options influences outcomes
6. Behavioural spillovers
• Theory:
• Behavioural changes in response to one policy spill over into other contexts
• Evidence for spatial spillovers but less for spillovers across policies
• Evidence
• Dikgang (2012)
• Convery, McDonell and Ferreira (2007)
• Relevance
• Spillovers from plastic bag tax enhanced responses to this policy
7. Hypothetical bias
• Theory:
• Respondents inflate their valuations in hypothetical scenarios when compared to
valuations in reality
• Therefore, estimates from hypothetical contexts are upwardly biased
• Evidence
• Johnston et al. (2017)
• Donaldson, Thomas and Torgerson (1997)
• Relevance
• Must incorporate multiple bias reducing/revealing measures.
8. Method
• Survey to gain public attitudes and valuations
• Linear and probabilistic regression analysis using survey data
• Hypothetical bias:
• Aim to mitigate in several ways including the question and answer designs
• Also aim to reveal the degree of bias in the estimates by using certainty
statements.
9. Initial engagement:
• Interviews:
• Discussion with University of Bath Estates team
• The ‘blue-planet’ effect – corroborates the spillover argument
• Incentives only work for specific goods
• Pilot study:
• Draft version of study sent to pilot group as pre-testing of survey design
• Feedback on language, design and timing.
• Academic language not amenable to wider public
10. Survey
• Part A: Descriptive
• Gender
• Age
• Employment
• Student stage
• Monthly income
• Monthly drink spend
• Part D: Valuations
• Reusable valuation
• Why not reusable
• Discount value
• Discount certainty
• 25p discount
• Tax value
• Tax certainty
• 25p tax
• Part B: Consumption
• Frequency
• Volume
• Availability
• Reusables
• Part C: Environmental concern
• Support latte levy
• Degree of support
• Support plastic bag tax
11. Survey design
• “The next section will ask you to report monetary values that you would be willing to pay if this were a
real-world experiment. This section is here to ask you to please consider your answers in the same context
as if you are spending your own money in a shop. As mentioned earlier, this survey is anonymous, so please
try to be as accurate as possible in your guesses as that will help my research here. Thanks again for your
honesty and help in answering this.”
• Q18) “What value of tax/discount would be sufficient for you to switch from disposable to reusable
containers?”
• A) 0p Would not switch, Price - (1-5p), Price - (6-10p), Price - (11-15p), Price - (16-20p), Price - (21-25p),
Price - (26-30p), Price - (31-35p), Price - (36-40p), Price - (41-45p), Price - (46-50p), Price - (51p+)
• Q19) “How certain are you about this valuation?”
• A) Definitely sure, Probably sure, Not sure.
• Q20) “Would a tax/discount of 25p be sufficient for you to exclusively switch to a reusable container?”
• A) Yes/No
12. Survey
• Sample:
• Respondents:
• 112
• Male/Female:
• 62/50
• Education:
• 76/112 in education
• Income:
• 94/112 income of £0-250 per month
• Modal tax valuation:
• Price + (21-25p)
• Modal discount valuation:
• Price – (51p+)
• Validity:
• Consistent with expectations
13. Research findings
Weak and inconsistent behavioural spillovers
Clear hypothetical bias influence on valuations
14. Statistics
• Descriptive statistics were sufficient to answer main research question
• Aimed to better understand the determinants of valuations
• OLS:
• Restrictions:
• Restricted: Only definitely sure responses
• Unrestricted: All responses included
• Specification:
• Valuation ~ Survey questions
• Narrow model: automatic model selection
• Expanded model: all survey questions
• Outcome:
• No significant variables in restricted, two in unrestricted
• Adequate AIC and R2 however
• Probit:
• Same restrictions and model selection as above
• Specification: likelihood of changing ~ Survey questions
• Report marginal effects for ease of inference
• More significant variables and more consistency between tax and discount
19. Research findings
• Plot of fitted vs residual
values from each OLS
estimation
• Can clearly see differences:
• Discount exceeds tax
• Unrestricted exceeds
`Definitely Sure’
• Supports first and third
hypothesis
20. Research findings
• Taxes more effective at lower cost than discounts
• Discount: mean £0.32, mode: £0.55
• Tax: mean: mean: £0.22 , mode: £0.25
• Behavioural spillovers weak
• No consistent inference possible.
• Generally for: Tax £0.18/Discount: £0.35
• Generally against: Tax: £0.16/Discount: £0.41
• Hypothetical bias mitigated but not eliminated
• The difference between tax and discount valuations fell with certainty
• Definitely sure: difference of £0.13
• Probably sure: difference of £0.11
• Not sure: difference of £0.04
21. Conclusions
• Results:
• Tax is more effective than a discount
• Behavioural spillovers exist, but weakly
• Hypothetical bias can be somewhat mitigated but not eliminated
• Implications:
• Policymakers must consider how the public perceives new policies
• Both in terms of framing effects and in the context of previous policies.
• Recommendations:
• Could scale the ‘Not-sure’ responses to be comparable to the ‘Definitely-Sure’ responses and re-
estimate.
• Could use model-fitting to improve regression analysis.
• Increased sample size and adding panel data element.
22. Any questions?
References:
• Convery, F., McDonell, S. and Ferreira, S., 2007. The
most popular tax in Europe? Lessons from the Irish
plastic bags levy. Environmental and resource
economics, 38(1), pp.1–11
• Poortinga, W., Whitmarsh, L. and Suffolk, C., 2013.
The introduction of a single-use carrier bag charge
in Wales: Attitude change and behavioural spillover
effects. ”Journal of Environmental Psychological., 36,
pp.240–247.
• Tversky, A. and Kahneman, D., 1981. The framing of
decisions and the psychology of choice. Science,
211(4481), pp.453–458
• Dikgang, J., 2012. Analysis of the plastic-bag levy in
South Africa. Resources Conservation and Recycling,
pp.59–65.
• Environmental Action Committee, 2018. Disposable
packaging: Coffee cups.
• Accessibility:
• Full text: bit.ly/Latte-Levy
• My research: bit.ly/KingPete
• Ethics:
• Request confirmed: 21/03
• Replication:
• Survey data and python data available upon
request
• Survey design available in full text