Table A1. Regression discontinuity...

Table A1.

Regression discontinuity estimation results testing the validity of alternative thresholds for scaffolding and earthwork projects through designated bidding only

	Estimation model 2 (first ordered)			Estimation model 3 (second ordered)
Cutoff	Estimated discontinuity (robust)			Estimated discontinuity (robust)
Cutoff	Coef.	Std. err.	P > z	Coef.	Std. err.	P > z
0.935	−0.038	0.172	0.826	−0.109	0.199	0.585
0.940	0.137	0.085	0.106	−0.127	0.120	0.291
0.945	0.259	0.077	0.001	0.350	0.085	0.000
0.950	−0.527	0.068	0.000	−0.515	0.069	0.000
0.955	−0.007	0.039	0.867	−0.104	0.049	0.033
0.960	−0.027	0.015	0.069	−0.021	0.018	0.255
0.965	0.015	0.020	0.447	−0.018	0.027	0.506

	Estimation model 2 (first ordered)			Estimation model 3 (second ordered)
Cutoff	Estimated discontinuity (robust)			Estimated discontinuity (robust)
Cutoff	Coef.	Std. err.	P > z	Coef.	Std. err.	P > z
0.935	−0.038	0.172	0.826	−0.109	0.199	0.585
0.940	0.137	0.085	0.106	−0.127	0.120	0.291
0.945	0.259	0.077	0.001	0.350	0.085	0.000
0.950	−0.527	0.068	0.000	−0.515	0.069	0.000
0.955	−0.007	0.039	0.867	−0.104	0.049	0.033
0.960	−0.027	0.015	0.069	−0.021	0.018	0.255
0.965	0.015	0.020	0.447	−0.018	0.027	0.506

Note(s):

This table presents the results of 14 RD estimations applied to scaffolding and earthwork projects conducted exclusively through designated bidding. The analysis tests for potential discontinuities in theproportion of winning bids at seven different threshold values. The estimation follows models 2 and 3 from Table 3 of the main text, which differ in polynomial order: model 2 uses a first-order (linear) regression, whereas model 3 adopts a second-order (quadratic) specification. Statistically significant discontinuities at the 10% level (based on the z-test) are shown in bold format. When estimated discontinuities are statistically significant in both models, the corresponding cells are shaded in light gray

Source(s): Author’s own work

[ViewLarge]

Sharing Unavailable