In my experience, it's more like "due to an uncontrollable urge that building technology in-house will be more fun, more interesting, more pleasing than using an existing solution" :)
...not that I'm completely innocent in this respect.
Indeed, retention is valuable.
> Some NIH will give your best staff something to really get their teeth into. A chance to create. A chance to contribute something original. An incentive to remain at your company.
Although it's not saying fun, pleasing, or interesting, it could be argued that this is what is meant, just with a more positive spin.
What's notably missing is whether the original, meaty creation provides value beyond retention (which a sibling comment points out does, indeed, have value).
Possible value to the staff is improved (or at least kept honed) skills, and possible value can go to the community for anything open-sourced. Of course, the latter can result in a proliferation of options that result in a "paralysis of choice" for which the article has an entire section.