Robots.txt does not prevent links from passing pagerank
We know this because sometimes pages that robots.txt
disallows end up ranking even when Googlebot can't crawl them. In those cases the "title" of the page that Google uses is usually based on the anchor text from the links. The pages rank based solely on the power of the external links.
I've also done experiments that show when internally linking to pages blocked by robots.txt
less pagerank is available to flow to other pages on the site.
Google can't tell what blocked pages link to
So even if blocked pages get some pagerank, they can't directly help the other pages on your site.
Google may re-assign pagerank to uncrawlable pages
Sites that block lots of pages with robots.txt
(and link to those pages internally frequently) seem to enjoy good rankings anyway. The same goes for sites that try to use nofollow
to sculpt pagerank. The practice doesn't seem to hurt as much as I would expect. This is likely because Google re-assigns "lost" pagerank to other pages on the site. Possibly by distributing it evenly across all indexed pages. Another explanation might be that Google has a concept of "domain authority" that they use in addition to Pagerank that helps in these cases.
Inbound links to blocked pages may help some
So my takeaway would be:
- All (non-spammy) inbound links (whether to blocked pages or not) should help some.
- Links to blocked pages are not going to help directly. They are not going to help specific pages rank. Nor is the anchor text going to matter much.
- Links to blocked pages probably don't help as much as links to crawlable pages (but it's hard to say how much less they help.)
So I would:
- Gladly accept inbound links to blocked pages.
- Not go out of my way to build links to blocked pages.
- Don't try to do anything stupid to claim every last ounce of link juice from blocked links. For example, you wouldn't want to make your site search results crawlable. That would be a direct violation of the Google webmaster guidelines.