Willingness-to-Pay for Energy Conservation and Free-Ridership on Subsidization – Evidence from Germany

We choose the conditional logit model as empirical point of departure, and ex-
plore the implications of re-estimating the model using three alternative discrete
choice models: the conditional logit model with error components, the random
parameters logit model, and error components layered over the random parame-
ters logit model, our most flexible model. The specification of utility in the most
general form is:

Uij = α + ( βι + Ui 1) Cij + ( β2 + Ui 2)∆ Qij

m∈z

+      ψhμjh + с,,

h∈{1,2}

where α is a constant that is specific to alternative j , C_ij is the investment
cost of household i for alternative j , and ∆Q_ij is the respective energy-savings
variable, computed as the difference in the building’s annual primary energy
demand in response to retrofit j.5 The vector z_i. = {income, energy consumption,
information access, east} contains the household-specific characteristics that enter
utility via interaction effects with investment cost and energy savings. Details
on data assembly for cost and energy savings are given in the appendix. Table
1 presents an overview of the data, including a listing of the 16 options and
the corresponding average costs and energy savings. The random parameters
βi1 =(β1 + Ui1) for investment cost and βi2 =(β2 + Ui2) for energy savings,
as well as the error components ψ_h ~ N(0,σψh),h ∈ {1, 2}, are only present
in the random-parameters and error components logit models, respectively. For
of auditors within this buffer. As a final step, we divided this count by the number of homes
(excluding apartment complexes) within the buffer. The variable thus created serves to capture
the relative availability of expert guidance on retrofits within the vicinity of the household.

⁵It is important to emphasize that because such savings accumulate over the lifetime of
the retrofit, the value of ∆Qij is not equivalent to the energy spot price of a MWh, but rather
will depend on several household-specific attributes, including time preference and expectations
about future energy prices.

12

<<   11   12   13   14   15   16   17   >>

More intriguing information