There are also studies that show otherwise, which the article - perhaps conveniently so - fails to mention:
https://www.gsb.stanford.edu/faculty-research/working-papers...
http://www.cosocloud.com/press-release/connectsolutions-surv...
https://www.zenefits.com/blog/remote-employees-happier-and-m...
There are extensive sets of studies that show all of that it generally doesn't, where it especially doesn't, and why it doesn't (which the article mentions.)
Data > anecdote.