You're only reading at surface level. It's described in the text, but for a quick sense of it look at the graphs on page 39 and the footnote - for earnings the participation is not considered but for hours worked and wages it only considers those in the labour force. You can see that the effect is less pronounced but it's still there. If I read this right women earn 10-15% less after having at least one child. Noticeably the dip in year 1 is much weaker, so in the overall earnings graph this dip might be due to women not working (much) in the first year and then going back to work - but on average with a comparably lower salary.
What you should also consider here is that this article is not looking at wage comparison within one company. It's looking at overall earnings in the population with a very thorough and deep analysis. It's an analysis for policy making, not for grandstanding. And for this whether women continue to work is also relevant - as the authors note this is quite a significant consideration even for labour force planning/statistics. Moreover thinking of e.g. pensions this is a very important point - as women will have lower pensions due to lower lifetime earnings.
Please do look at the whole picture, rather than just try to validate your viewpoint. There are many interesting tidbits here, for example that earnings fall 10% for each following child. The graph on p42 comparing childless women and mothers is also particularly impressive.