Table 2: |
Estimation Results of Logit Models | |||
CLogit |
RP. Logit | |||
×10-2 |
CLogit |
RP. Logit |
with EC. |
with EC. |
Cost (Cij) |
-12.667** |
-12.873** |
-18.469** |
-18.410** |
Energy Savings (∆Qij) |
22.700** |
23.296** |
32.806** |
32.843** |
Interaction of Cost with | ||||
Energy Consumption |
0.101 ** |
0.102** |
0.157** |
0.157** |
Income |
-0.124* |
-0.129* |
-0.177** |
-0.182** |
Information Access |
0.116 |
0.121 |
0.157* |
0.157* |
East |
4.783** |
4.946** |
7.723** |
7.724** |
Interaction of Energy Savings with | ||||
Energy Consumption |
-0.162** |
-0.167** |
-0.239** |
-0.241** |
Income |
0.100 |
0.105 |
0.152* |
0.158* |
Information Access |
-0.020 |
-0.022 |
-0.020 |
-0.020 |
East |
-0.980 |
-1.073 |
-2.817 |
-2.770 |
Standard deviation for random parameters distribution | ||||
Cost |
4.732 |
0.110 | ||
Energy Savings |
1.289 |
0.318 | ||
Standard deviation for error components | ||||
No renovation |
0.365 |
0.501 | ||
Annoying renovation |
210.741** |
211.819** | ||
Log-Likelihood |
-5054 |
-5054 |
-5035 |
-5034 |
**significant at the 1% level, |
* significant at the 5% level. Alternative specific constants not presented. |
3 Results
In this section we present the results of the discrete choice models. The section
begins with a cataloging of the coefficient estimates followed by a comparison of
model fit. Thereafter we derive the marginal WTP and present its distribution
over households. The section closes with a discussion of free-ridership and policy
implications.
3.1 Coefficient Estimates and Model Fit
Table 2 presents the results of a conditional-, random parameters-, error com-
ponents, and error components with random parameters logit model. All four
models tell a consistent story. The signs of the significant coefficients are the
15