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

Table 1: Mean Investment Cost and Mean Energy Savings

example, if u_i1 = u_i2 = 0 for all households i, and σψ² ₁ = σψ²₂ = 0, then Equation
(8) collapses to the conditional logit specification.

In specifying the error components ψh , the aim was to capture latent effects
specific to an outcome or a set of outcomes. We explored several alternatives,
guided by the considerations noted above concerning both the hidden costs and
benefits of, respectively, grime and prestige associated with particular retrofit
options. The presented specification follows closely Cameron’s (1985) nested logit
analysis by incorporating two error components, the first of which distinguishes
the binary decision concerning whether to retrofit, and the second of which groups
13 of the remaining retrofit combinations that tend to produce annoying levels
of dirt and disarray (indicated in the final column of Table 1). We also explored
models with additional error components for alternatives conferring prestige, but

13

<<   <   12   13   14   15   16   17   18   >   >>