. 2020 Jul 16;66:101887. doi: 10.1016/j.labeco.2020.101887

Table 3.

Subgroup analysis for single-adult households

	Test I.1	Test I.2	Test II.1	Test II.2
Transition effects, full sample:
Wave controls (1^st / 6^th specification)	0.498^⁎⁎*	-0.440^⁎⁎*	-0.239^⁎⁎	0.318^⁎⁎*
	(0.089)	(0.155)	(0.099)	(0.083)
Full set of controls (3^rd / 8^th specification)	0.435^⁎⁎*	-0.416^⁎⁎*	-0.228^⁎⁎	0.242^⁎⁎*
	(0.090)	(0.151)	(0.103)	(0.082)
Number of observations	12,514	1,689	23,112	2,163
Number of ‘treated’ observations	411	188	345	681
Transition effects, single-household sample:
Wave controls (1^st / 6^th specification)	0.372^⁎⁎*	-0.517^⁎⁎	-0.340^⁎⁎	0.253^⁎⁎
	(0.124)	(0.218)	(0.141)	(0.122)
Full set of controls (3^rd / 8^th specification)	0.286^⁎⁎	-0.434*	-0.261*	0.213*
	(0.127)	(0.223)	(0.145)	(0.128)
Number of observations	7,226	782	6,285	962
Number of ‘treated’ observations	210	86	143	274

Note. *denotes significance at the 10% level

^⁎⁎

at the 5% level and ^⁎⁎⁎at the 1% level. Robust standard errors in parentheses. For the upper panel see the notes of tables 1 and 2. The lower panel reports transition effects of tests I.1, I.2, II.1, II.2 estimated for single-adult households only, i.e. one single adult or single parent with the same number of dependent children at both reference points in time. Entropy balancing was applied on the first moment. The conditioning variables are the same as for the full sample, except for variables that do not vary across single households (number of adults in household, marital status). Controls (wave, full set) refer to respective columns of tables 1 and 2. Unlike the main results, the full set of controls does not include the changes in household size.

Source. PASS 2007-2018.