Academia is mostly about signalling, not entirely, but mostly (see Robin Hanson's http://www.overcomingbias.com posts on academia). Making it difficult and expensive to publish increases the signalling effect and helps keep out the riff-raff.