A key insight from the scientific literature on climate change is that the projected impacts of climate
change are not expected to be equally distributed among nations. Developing regions are expected
to bear a larger share of the adverse impacts, compared to developed regions. This will have
disproportionate effects, as according to the theory of declining marginal utility of income,
individuals living in poor regions will suffer a larger utility loss from a dollar worth of climate
damage relative to individuals living in rich regions who are subject to the same one dollar worth of
climate damage. This concern can be integrated into the central policy assessment tool in the
context of climate policy, which is the Social Cost of Carbon (SCC). The SCC captures the
economic costs to societies that a marginal increase in greenhouse gas emissions today will have in
the future and is calculated by integrated assessment models, typically under the assumption of a
global policy maker. Equity weights, also called distributional weights can be included in the SCC
in order to give more weight to damages that befall poorer individuals, relative to those felt by more
well-off individuals. Given the current limitations of international cooperation on climate change,
the chances of having one body/actor to direct climate policy at the global level seem dim -
employing a national outlook and operating from the perspective of a national policymaker thus
becomes relevant. A national policymaker who wishes to account for distributional outcomes
abroad can specify the SCC to include other-regarding, also called social preferences.
The objective of this thesis is to explore the empirical foundation of accounting for distributional
social preferences in the context of climate policy. This is a relevant question to pursue for three
reasons: 1) the empirical existence of distributional social preferences in the context of climate
policy has not previously been explored, 2) the theoretical implications of a policymaker with
distributional social preferences, is a higher price on carbon emissions, which ceteris paribus will
generate grounds for a more aggressive level of climate policy, and finally 3) the efficiency of
climate policies requires a design founded on the actual preference structure of individuals. A
climate policy that reflects the general public’s preferences towards distributional outcomes is more
likely to be accepted and thereby more efficient in attaining its goal.
This PhD thesis consists of five papers that explore distributional social preferences in the context
of climate policy and is based on three separate data collections. The first data collection is based
on a standard discrete choice experiment (DCE) using tax as payment vehicle and includes a split to
test the effect of valence-based framing. Papers 1, 3 and 5 are based solely on the first data
collection. The second data collection, which Paper 2 is based on, involves an incentivized DCE
using a real donation mechanism, while the third data collection is based on a standard DCE using a
donation as a payment vehicle. The analysis in Paper 4 is based on all three data collections.
Paper 1 presents a theoretically based characterization of distributional social preferences into two
proposed behavioural hypotheses and finds that a small majority of Danes exhibit distributional
social preferences reflecting inequity aversion, a finding which has been tested robust against
several sensitivity checks. The paper also presents evidence that a non-negligible share of
individuals appears to signal a commitment to the climate cause, and that a small share of
respondents prefers not to support climate policies at all. Paper 2 explores the characterization of
distributional social preferences in an experimental context where individuals make actual
donations towards CO2 mitigation and adaptation projects. The results of this paper suggest an
average tendency for individuals to hold distributional social preferences, which are likely to be
heterogeneous in nature. Paper 3 investigates whether the systematic part of distributional
preference heterogeneity can be linked to two behavioural drivers, 1) the individual’s expectation
regarding the degree of climate change and 2) individual time preferences. The findings of the
paper suggest that both drivers can be used to further the understanding of distributional social
preferences in the context of climate policy. The focus of Paper 4 is to explore how changes in
survey design influence the conclusions reached on individuals’ choice of climate policy. The
findings indicate that elicitation format, bid range and type of payment vehicle can be potential
drivers behind the observed differences in both preference structures and welfare estimates for
climate policy. Finally, Paper 5 investigates how valence-based attribute level framing influences
the observed preference structure and welfare estimates for climate policy. The results suggest that
framing significantly influences the observed preference structure and the welfare estimates reveal a
significant spill-over effect on an unframed policy attribute.
Taken together, this thesis suggests that a substantial share of the population considers the
distributional outcomes of climate policy to be of importance – a finding which is consistent across
the different experimental contexts. Moreover, these distributional concerns are likely to be
heterogeneous in nature and relate to the individual’s expectation regarding the degree of climate
change as well as individual’s time preferences. This thesis also contributes to the literature on
stated preference design, by confirming that several survey design features influence the observed
inference from choice models in the context of a moral public good. In particular, this thesis finds
that framing climate policies in terms of gains or losses do not influence the role of distributional
concerns, a framing that highlights loss, however, overall leads to a higher willingness to support
climate policies. This thesis presents the following implications for future climate policy design: 1)
it provides an argument for the inclusion of distributional concerns in climate policy design; 2)
policymakers can enhance the support for climate policies by stressing the presence of co-benefits;
3) loss framing will increase the support to climate policies, and finally; 4) policymakers should
recognise that support for climate policies will vary between different segments of the populatio