Estimation of standard errors and treatment effects in empirical economics—methods and applications

Journal for Labour Market Research

Table 13 OLS estimates of Cobb-Douglas functions with company-level pact dummy (CLP) as regressor, decreasing shares of n(CLP=1)/n; dependent variable: logarithm of sales

	\(\hat{\beta}_{\mathrm{CLP}}\)	Std.err.	t
\(\overline{\mathrm{CLP}}=\boldsymbol{0.0693}\)	0.1231	0.0236	5.22
\(\overline{\mathrm{CLP}}=0.0624\)	0.1209	0.0246	4.92
\(\overline{\mathrm{CLP}}=0.0533\)	0.1299	0.0259	5.02
\(\overline{\mathrm{CLP}}=0.0477\)	0.1131	0.0275	4.11
\(\overline{\mathrm{CLP}}=0.0407\)	0.1006	0.0295	3.41
\(\overline{\mathrm{CLP}}=0.0336 \)	0.1005	0.0322	3.12
\(\overline{\mathrm{CLP}}=0.0273\)	0.1429	0.0356	4.01
\(\overline{\mathrm{CLP}}=0.0207\)	0.1446	0.0403	3.39
\(\overline{\mathrm{CLP}}=0.0135\)	0.1357	0.0486	2.79
\(\overline{\mathrm{CLP}}=0.0067\)	0.1887	0.0671	2.80

Note: IAB Establishment Panel 2006–2010; n=31,985. In the first line the estimation with the original sample and \(\overline{\mathrm{CLP}}=0.0693 \) is presented. Next, only 90 % of the firms with CLP=1, where \(\overline{\mathrm{CLP}} =0.0624\), are considered. The random selection of the CLP firms is based on a rectangular distribution of the CLP firms. The determination of the following lines is analogous to that of the second line