A Carbon Price by Another Name May Seem Sweeter David Hardisty1 Alec Beal1, Ruben Lubowski2, Annie Petsonk2, Rainer Romero-Canyas2 1 The University of British Columbia 2 The Environmental Defense Fund 1

Origins

Carbon Pricing in Aviation Some countries have voluntary carbon pricing, others don’t (ICAO; CORSIA) Goal: Design effective and accepted carbon pricing policies Where to set the “point of regulation”? A critical goal of EDF’s new 5-year strategic plan is to ensure that by 2020, one quarter of the world’s carbon dioxide emissions, including those from international aviation, are covered by durable, declining limits achieved with a carbon price, and that these are on track to having one half of emissions covered by 2030. Designing emissions trading systems and other carbon pricing policies that are effective and appropriate is thus a critical policy challenge. Through our research we hope to ensure that these carbon pricing policies not only make a significant impact on climate change, but are also accepted amicably by the general public. One fundamental question in the design of effective carbon pricing policies is where to set the “point of regulation” at which either organizations or consumers face compliance responsibilities for paying a price on emissions 3

Points of Regulation “Upstream” – Producers and importers “Downstream” – Companies & Customers Policy makers have various options for choosing a point of regulation in the case of fossil fuel emissions where emissions are directly proportional to the carbon content of the fuel. -They may choose to set an upstream point of regulation such as when compliance responsibilities fall on the miners, producers, and importers of carbon. -Or they may choose to set a downstream point of regulation such as when compliance responsibilities fall on the consumer. -In theory, producers and importers regulated upstream will pass the carbon price through to consumers, sending a price signal which would increase the cost of carbon-consuming goods and services, thus reducing demand. Similarly, consumers facing a downstream carbon price may respond by reducing their demand; producers will then see the same reduction in profits as if they had faced the carbon price upstream and correspondingly passed on the cost to consumers. So theoretically, setting the point of regulation upstream or downstream should have similar economic effects for all parties involved. -Thus given their theoretical equivalence, an important, previously unexplored question becomes, which “stream” do consumers prefer? Put differently, would consumers’ willingness to pay an additional carbon fee differ if it was labelled as upstream versus downstream? Current thinking among policy makers is that a downstream point of regulation may make the price increase more salient to consumers, however, this claim is largely anecdotal. Similar economic impacts for “upstream” and “downstream” What are the psychological impacts on consumers? (Aldy & Pizer, 2009) 4

25% FAT 75% MEAT Attribute Framing Labelling and phrasing can impact attitudes, behavior 25% FAT 75% MEAT I’ll now talk about some background research which informed this work starting with some work on attribute framing. Attribute framing refers to specific labelling, wording, phrasing that can impact how one perceives something. Attribute framing can be easily understood with an example from previous literature. In a classic study, consumers were given a delicious hamburger, tasted it and were asked how much they would be willing to pay for it. Consumers were willing to pay more for a burger labeled as -75% meat than one labeled -25% fat, even though it was the exact same hamburger (Levin & Gaeth, 1988) 5

(Hardisty, Johnson & Weber, 2010; Sussman & Olivola, 2011) Carbon Fee Labelling “Tax” labelling less popular than other frames (such as “offset”) Particularly among anti-tax political parties such as Republicans A similar principle guides research from Dave Hardisty and others in the pro-environmental domain. - Some of their previous work has shown that costs labeled as "taxes" are more odious to consumers than other equivalent financial costs, and this is particularly true for political conservatives - Specifically, when choosing between otherwise identical products or services, where one option included a surcharge for emitted carbon dioxide, Americans were more likely to prefer to purchase products and services when the environmental cost was labelled as an offset rather than a tax. (Hardisty, Johnson & Weber, 2010; Sussman & Olivola, 2011) 6

The Current Research Test consumer preferences in the aviation industry in 7 studies “Upstream” vs. “Downstream” “Tax” vs. “Offset” Alternative Hypothesis 1: Consumers are most accepting of an “Upstream Tax” or a “Downstream Offset” Bad things for them Good things for me Alternative Hypothesis 1: Consumers are most accepting of a carbon price when labelled as an “Upstream Offset” Cause: Hold the polluters responsible Consequence: Be good to the environment In a series of three studies, the current research tests consumer perceptions of an airplane carbon fee under several different labels. Specifically, we test whether participants’ willingness to pay an additional carbon fee on their airplane flights will vary if that fee is described as - Upstream vs. Downstream. - And/or will vary if that fee is described as a tax or an offset. Based on the idea that an upstream point of regulation will make the price increase less salient, and previous research outside of the aviation industry suggesting that consumers are more accepting of fees labelled as offsets. We hypothesized that consumers may be most accepting of a carbon fee labelled as an Upstream Offset. The work I’m going to be talking about today is available in more detail in our working paper which can be found online and is currently one of the top-ten most downloaded papers on the SSRN framework. 7

Study 1: Method 318 U.S. MTurkers (54% men, M age = 35) Six conditions, between subjects Four experimental Two controls In Study 1 a large sample of participants were recruited through Amazon Mechanical Turk and participated online. -Participants were randomly assigned to one of eight conditions 8

Study 1: Method Four experimental conditions: Upstream Tax, Upstream Offset Downstream Tax, Downstream Offset Brief regulatory policy description Immediately following the manipulation, participants in the four experimental conditions gave responses on our main dependent variable. -Which was a gauge of their willingness to purchase pro-environmental flights Flight pair preferences 9

Study 0: Method [Upstream Offset] As you may know, carbon dioxide emissions are produced by many human activities, such as burning fossil fuels for driving, generating electricity, or flying, and these emissions are contributing to climate change. Some governments are considering imposing a mandatory carbon offset program for aviation fuel production and import. Aviation fuel producers and importers would need to purchase offsetting emission reductions (e.g., from programs that reduce emissions at electric power plants, or that protect forests which would otherwise be chopped down). “Offsetting” would send a price signal that should, over time, elicit a market response across the entire aviation industry. It’s estimated that the program would add $5.70 to the cost of a flight from New York to Los Angeles. Here is the Upstream Offset regulatory policy for example which is slightly lengthier and describes in a bit more detail how offsetting programs work and from where offsetting emission reductions are purchased. 10

Study 1: Method Upstream Offset Some governments are considering imposing a mandatory carbon offset program for aviation fuel production and import. Aviation fuel producers and importers would need to purchase offsetting emission reductions to address a share of the associated greenhouse gas emissions. Each of the shorter policy descriptions in Study 2 was specifically designed to be matched as closely as possible in length and general content. While each being under 40 words in length. 11

Study 1: Method Downstream Offset Some governments are considering imposing a mandatory carbon offset program for airplane travel and cargo. Airline customers would need to purchase offsetting emission reductions to address a share of the associated greenhouse gas emissions. Between conditions, the only text that varied was the stream. Downstream seen here in green. And whether the label was a tax or an offset. Tax seen here in blue. 12

Study 1: Method Upstream Tax Some governments are considering imposing a mandatory carbon tax program for aviation fuel production and import. Aviation fuel producers and importers would need to pay a tax to address a share of the associated greenhouse gas emissions. Special care was taken to make sure the policy descriptions in study 2 contained as little superfluous information as possible. With the hopes that these policy descriptions may be actionable in a real-world setting. 13

Study 1: Method Downstream Tax Some governments are considering imposing a mandatory carbon tax program for airplane travel and cargo. Airline customers would need to pay a tax to address a share of the associated greenhouse gas emissions. Special care was taken to make sure the policy descriptions in study 2 contained as little superfluous information as possible. With the hopes that these policy descriptions may be actionable in a real-world setting. 14

plus an additional $14.00 carbon tax on airplane travel. Study 1: Method Downstream Tax Specifically, participants were asked to imagine that they were planning a vacation, and were presented with two pairs of similarly priced flights to little known vacation destinations for each pair of flight decisions. In all cases Flight B included an additional $14.00 carbon fee whose description varied in stream and frame depending on condition. For example, one flight pair decision in the Downstream Tax condition had participants choose between “Flight A to the Island of Tortola for $625.25 OR Flight B to the Island of Anegada for $605.25 plus an additional $14.00 carbon tax on airplane travel.” -Importantly, we weren’t really interested in people’s preference for Flight B over Flight A at baseline, we were interested in how this preference DIFFERED between condition. So in all conditions, Flight A was exactly the same and what we varied was the description of Flight B. Price: $322.99 Price: $310.99 plus an additional $14.00 carbon tax on airplane travel. 15

Study 1: Method Upstream Tax Price: $322.99 Price: $310.99 Specifically, participants were asked to imagine that they were planning a vacation, and were presented with two pairs of similarly priced flights to little known vacation destinations for each pair of flight decisions. In all cases Flight B included an additional $14.00 carbon fee whose description varied in stream and frame depending on condition. For example, one flight pair decision in the Downstream Tax condition had participants choose between “Flight A to the Island of Tortola for $625.25 OR Flight B to the Island of Anegada for $605.25 plus an additional $14.00 carbon tax on airplane travel.” -Importantly, we weren’t really interested in people’s preference for Flight B over Flight A at baseline, we were interested in how this preference DIFFERED between condition. So in all conditions, Flight A was exactly the same and what we varied was the description of Flight B. Price: $322.99 Price: $310.99 plus an additional $14.00 carbon tax on aviation fuel production and importation. 16

plus an additional $14.00 carbon offset for airplane travel. Study 1: Method Downstream Offset Specifically, participants were asked to imagine that they were planning a vacation, and were presented with two pairs of similarly priced flights to little known vacation destinations for each pair of flight decisions. In all cases Flight B included an additional $14.00 carbon fee whose description varied in stream and frame depending on condition. For example, one flight pair decision in the Downstream Tax condition had participants choose between “Flight A to the Island of Tortola for $625.25 OR Flight B to the Island of Anegada for $605.25 plus an additional $14.00 carbon tax on airplane travel.” -Importantly, we weren’t really interested in people’s preference for Flight B over Flight A at baseline, we were interested in how this preference DIFFERED between condition. So in all conditions, Flight A was exactly the same and what we varied was the description of Flight B. Price: $322.99 Price: $310.99 plus an additional $14.00 carbon offset for airplane travel. 17

Study 1: Method Upstream Offset Price: $322.99 Price: $310.99 Specifically, participants were asked to imagine that they were planning a vacation, and were presented with two pairs of similarly priced flights to little known vacation destinations for each pair of flight decisions. In all cases Flight B included an additional $14.00 carbon fee whose description varied in stream and frame depending on condition. For example, one flight pair decision in the Downstream Tax condition had participants choose between “Flight A to the Island of Tortola for $625.25 OR Flight B to the Island of Anegada for $605.25 plus an additional $14.00 carbon tax on airplane travel.” -Importantly, we weren’t really interested in people’s preference for Flight B over Flight A at baseline, we were interested in how this preference DIFFERED between condition. So in all conditions, Flight A was exactly the same and what we varied was the description of Flight B. Price: $322.99 Price: $310.99 plus an additional $14.00 carbon offset for aviation fuel production and importation. 18

plus an additional $14.00 carbon offset for airplane travel. Study 1: Method Price: $322.99 Price: $310.99 plus an additional $14.00 carbon offset for airplane travel. For each of the two flight pairs participants indicated the likelihood that they would purchase flight B (the flight which included the carbon fee) instead of flight A on scale of 1 – definitely not, to 7- definitely. We aggregated scores on this item for each of the two flight pair choices in study one, and this composite measure was our main dependent variable indicating a willingness to purchase pro-environmental flights. 19

Study 1: Method Two control conditions: No-fee control No-information control Immediately following the manipulation, participants in the four experimental conditions gave responses on our main dependent variable. -Which was a gauge of their willingness to purchase pro-environmental flights Flight pair choices 20

Study 1: Method No-fee control Price: $322.99 Price: $310.99 Specifically, participants were asked to imagine that they were planning a vacation, and were presented with two pairs of similarly priced flights to little known vacation destinations for each pair of flight decisions. In all cases Flight B included an additional $14.00 carbon fee whose description varied in stream and frame depending on condition. For example, one flight pair decision in the Downstream Tax condition had participants choose between “Flight A to the Island of Tortola for $625.25 OR Flight B to the Island of Anegada for $605.25 plus an additional $14.00 carbon tax on airplane travel.” -Importantly, we weren’t really interested in people’s preference for Flight B over Flight A at baseline, we were interested in how this preference DIFFERED between condition. So in all conditions, Flight A was exactly the same and what we varied was the description of Flight B. Price: $322.99 Price: $310.99 21

No-information control Study 1: Method No-information control Specifically, participants were asked to imagine that they were planning a vacation, and were presented with two pairs of similarly priced flights to little known vacation destinations for each pair of flight decisions. In all cases Flight B included an additional $14.00 carbon fee whose description varied in stream and frame depending on condition. For example, one flight pair decision in the Downstream Tax condition had participants choose between “Flight A to the Island of Tortola for $625.25 OR Flight B to the Island of Anegada for $605.25 plus an additional $14.00 carbon tax on airplane travel.” -Importantly, we weren’t really interested in people’s preference for Flight B over Flight A at baseline, we were interested in how this preference DIFFERED between condition. So in all conditions, Flight A was exactly the same and what we varied was the description of Flight B. Price: $322.99 Price: $324.99 22

Flight Pair Choices Island of Tortola $625.25 Island of Anegada $605.25 Isla Guamblin $1,156.16 Isla Melchor $1,146.16 Rarotongo $772.50 Mo’orea $764.50 Island of Nevis $322.99 Island of Bequia $310.99 Majuro $1,025.00 Upolu $1,011.00

Study 1: Design – 2x2+2? Four experimental conditions: Upstream Tax, Upstream Offset Downstream Tax, Downstream Offset Two control conditions (did not read a regulatory policy): No-Information Control, No-Fee Control Hypothesis 1: Consumers are most accepting of an “Upstream Tax” or a “Downstream Offset” Alternative Hypothesis 1: Consumers are most accepting of a carbon price when labelled as an “Upstream Offset” Finally, participants in all six conditions completed -A demographic questionnaire which assessed things like sex, age, income, and political orientation. In our paper we do some additional analyses with these variables and uncover some potential differences in effects between democrats and republicans, but in the interest of time I won’t go into those analyses here. 24

Does Carbon Fee Labelling Affect Willingness to Purchase Pro-Environmental Flights? Is the shorter upstream offset labelling more effective than other shorter policies which label a carbon fee differently? -It appears so! Preference for flights carrying a $14.00 carbon fee in the Shorter Upstream Offset condition was significantly greater when compared to the Shorter Upstream Tax condition, Shorter Downstream Offset condition, and Shorter Downstream Tax condition. This preference in the Shorter Upstream Offset condition was also significantly greater when compared to the No-Information Control condition in which participants did not read a proposed regulatory policy and no information was given regarding the carbon fee. Finally, individuals in the Upstream Offset condition were marginally more likely to prefer flights carrying a carbon fee when compared to the No-Fee Control condition in which the $14.00 carbon fee was not even applied. . 25

Upstream Offset policy description Study 1: Results “Upstream Offset” label most effective at eliciting pro-environmental flight preference More effective than other labels More effective than no label of the fee at all Marginally more preferred than less expensive flights to same destination (!) In Study 1 individuals who had read a brief description of a proposed regulation for a carbon fee described as a “carbon offset on aviation fuel production and importation” consistently reported a greater preference to purchase flights carrying a $14.00 carbon fee (versus similarly priced flights with no carbon fee) than, -individuals who read other brief descriptions describing the carbon fee under different frames and/or at a different stream -individuals who did not read any description of an additional carbon fee and for whom this $14.00 fee was embedded in the cost of the flight, and -individuals who did not read any description of an additional carbon fee and for whom this $14.00 fee was not even applied to the cost of the flight. Upstream Offset policy description “… plus an additional $14.00 carbon offset for aviation fuel production and importation.” 26

Open questions What about support for regulation? Is a “mandatory” offset still appealing? Why is upstream offset so effective? Based on qualitative studies, consumers: want to help the environment want the responsible party to be regulated Hypothesis 2: “upstream offset” is perceived to address both the cause and the consequences of emissions

Study 2: Method 1215 U.S. MTurkers (47% men, M age = 37) Three conditions Upstream Offset, Downstream Offset , Upstream Tax In Study 3, a large sample of participants was recruited through Amazon Mechanical Turk and participated online. Participants were randomly assigned to one of three conditions Policy description Flight pair prefs 28

Study 2: Method Support for policy: “How much would you support the implementation of this carbon regulatory program on a national level?” Impact: “Would you expect this carbon regulatory program to have a significant impact on climate change?” Accountability: “Does this carbon regulatory program ensure that those who are responsible for carbon emission pollution are the ones who pay to reduce it?” Policy description Policy description Policy questions Flight pair prefs 29

Does Carbon Fee Labelling Affect Willingness to Purchase Pro-Environmental Flight? Did we replicate the findings of our previous 2 studies? The “upstream offset” condition inspired greater pro-environmental flight preference than the “upstream tax” and marginally greater pro-environmental flight preference than the “downstream offset” conditions; t(796) = 3.00, d = 0.21, p = .003, and t(835) = 1.90, d = 0.13, p = .058, respectively. 30

Does Carbon Fee Labelling Affect Support for Implementation at a National Level? Did participants support of the implementation of the carbon regulatory program on a national level differ by condition? Policy support in the “upstream offset” condition was marginally greater when compared to the “upstream tax”, t(796) = 1.79, d = 0.13, p = .073, and significantly greater when compared to the “downstream offset” conditions, t(835) = 3.46, d = 0.24, p = .001. 31

Why?

Does Carbon Fee Labelling Affect Perceived Environmental Impact? Did participants’ expectations that the carbon regulatory program would have a significant impact on climate change differ between condition? Yes. Perceived environmental impact in the “upstream offset” condition was significantly greater when compared to the “upstream tax”, t(796) = 2.79, d = 0.20, p = .005, and the “downstream offset” conditions, t(835) = 2.70, d = 0.18, p = .007. 33

Does Carbon Fee Labelling Affect Perceived Carbon Emissions Accountability? Did participants’ expectations that the carbon regulatory program would ensure that those who are responsible for carbon emission pollution are the ones who pay to reduce it differ by condition? Yes. Perceived carbon emissions accountability in the “upstream offset” condition was not statistically different when compared to the “upstream tax” condition, t(796) = 1.27, d = 0.09, p = .21, but was significantly greater when compared to the “downstream offset” condition, t(835) = 4.14, d = 0.288, p < .001. 34

Study 2: Summary Upstream Offset policy description (vs. other descriptions): Greater pro-environmental flight preference Greater policy support Greater perceived environmental impact Greater perceived carbon emissions accountability Mediation? In order to get at the “why” question, we wanted to see whether these latter two perceptions of the upstream offset policy were actually driving pro-environmental flight preference and/or policy support. To answer these questions we turned to mediation analyses. 35

Mediation of Pro-Environmental Flight Choice? First, focussing only on the two “upstream” conditions, we tested whether perceived environmental impact and/or perceived carbon emissions accountability mediated the relationship between condition (coded -1= upstream tax and 1= upstream offset) and pro-environmental flight preference. Results showed that condition was related to pro-environmental flight preference indirectly through perceived environmental impact, indirect effect = 0.07 (SE = 0.03), 95% CI = [.02, .13], but not perceived carbon emissions accountability, indirect effect = 0.02 (SE = 0.01), 95% CI = [-.007, .05]. With the 5-flight pair choice composite as a measure of pro-environmental flight preference, this indirect effect was marginally significant (among the subsample for whom this analysis was possible), indirect effect = 0.07 (SE = 0.03), 95% CI = [-.02, .17]. Indirect effect of Environmental Impact = 0.07* (SE = 0.03), 95% CI = [.022, .13] Indirect effect of Emissions Accountability = 0.02 (SE = 0.01), 95% CI = [-.007, .05] 36

Mediation of Pro-Environmental Flight Choice? Next, focussing only on the two “offset” conditions (coded -1= downstream offset and 1= upstream offset), results showed that condition was related to pro-environmental flight preference indirectly through perceived environmental impact, indirect effect = 0.07 (SE = 0.02), 95% CI = [.02, .12], and perceived carbon emissions accountability, indirect effect = 0.04 (SE = 0.02), 95% CI = [.02, .08]. Indirect effect of Environmental Impact = 0.07* (SE = 0.02), 95% CI = [.02, .12] Indirect effect of Emissions Accountability = 0.04* (SE = 0.02), 95% CI = [.02, .08] 37

Mediation of Policy Support? Next, we tested whether perceived environmental impact and/or perceived carbon emissions accountability mediated the relationship between condition and policy support, focussing first on the two “upstream” conditions (coded -1= upstream tax and 1= upstream offset). Results showed that condition was related to policy indirectly through perceived environmental impact, indirect effect = 0.11 (SE = 0.04), 95% CI = [.04, .19], but not perceived carbon emissions accountability, indirect effect = 0.02 (SE = 0.01), 95% CI = [-.009, .04]. Indirect effect of Environmental Impact = 0.11* (SE = 0.04), 95% CI = [.04, .19] Indirect effect of Emissions Accountability = 0.02 (SE = 0.01), 95% CI = [-.009, .04] 38

Mediation of Policy Support? Next, focussing only on the two “offset” conditions (coded -1= downstream offset and 1= upstream offset), results showed that condition was related to policy support indirectly through perceived environmental impact, indirect effect = 0.10 (SE = 0.04), 95% CI = [.02, .17], and perceived carbon emissions accountability, indirect effect = 0.07 (SE = 0.02), 95% CI = [.03, .11]. Indirect effect of Environmental Impact = 0.10* (SE = 0.04), 95% CI = [.02, .17] Indirect effect of Emissions Accountability = 0.07* (SE = 0.02), 95% CI = [.03, .11] 39

Study 2: Summary Upstream offset condition → greater pro-environmental flight choice and policy support Mediated by perceptions of impact and accountability Brief regulatory policy description Brief regulatory policy description Policy-specific questions Flight pair choice 40

Limitations and Future Directions How culturally specific is this? Indian replication study (and Brazil on the way) However, all of our participants were Americans and more work is needed before these policy recommendations can be proffered to countries outside of the United States. -We are actually in the midst of conducting a replication of Study 5 with Indian participants. -We’ve had to make a few modifications like converting the cost of flights into rupees but this study which is currently running should help shed light on whether these findings generalize to other English-speaking countries. In the future we may wish to explore which labels are successful in non-English speaking nations as well. Price: ₹10,629 Price: ₹10,289 plus an additional ₹238 carbon offset for aviation fuel production and importation. 41

Limitations and Future Directions Field data! Reducing consumption (plastic bag tax) Pro-social consumption: Fair trade Charitable contributions Conferences??? Another important consideration is whether a difference in effects would be observed between consumers from different political parties. -While Republicans may respond negatively to the "downstream tax" framing, Democrats may not, which would be consistent with the results of previous research. I didn’t have time to talk about the possible generalizability across political parties but we do see some preliminary evidence of this. -Also, other product domains should also be explored. In particular, while consumer leisure air travel may be seen as a relatively discretionary expenditure, and therefore an acceptable target for carbon regulation, other product domains -Automotive travel, or home electricity may be less seen as less discretionary, and consumers may therefore have greater concerns about carbon regulation in these domains. 42

100% GOOD 0% BAD Conclusions Upstream offset framing may improve consumer acceptance of carbon pricing in the aviation industry Upstream/downstream framing is a promising nudge to test in other consumer contexts 100% GOOD 0% BAD I encourage anyone who is interested to check out our frequently updated working paper describing these findings which is currently one of the top-ten most downloaded papers on the SSRN framework. -The current findings have major real-world policy implications for the new market-based measure adopted for the international civil aviation sector, which has an offsetting program as a central feature. Our research suggests that airlines might wish to highlight their carbon offset purchases for their customers in order to increase customer receptivity. -If the effectiveness of upstream offset regulatory policy labelling is found to generalize beyond the aviation sector to other domains, then these findings could have other important implications for other pro-environmental behavior more broadly. -And that would be 100% good; -0% bad; and win however you want to frame it. 43

Thank you! 44