Of course voice will be faster--eight out of ten times at least one involved party's main goal is just to get off the phone.
Optimizing for time is a mistake. Optimize for disruption (I can switch back to other work while waiting for a chat reply--there's context-switching overhead there, but not nearly as much as a phone call), retention (I can refer back to what someone said in chat, rather than "hey, Jane, what was that hostname again?", waiting for an atrocious speech-to-text conversion, or manually seeking around in a call recording), and mutual investment in communication (professional or not, a synchronous call usually has at least one participant that wants to be there less than the other[s]--in chat, people can participate when they're most interested in doing so, meaning that their contributions are less half-assed).
There are situations where sync is best (trying to pitch something or persuade colleagues of something is a big one--it's just easier to reject/be reluctant without engaging critically via chat for some reason). But a lot of the time chat is more efficient.
I've found that many people I've worked with who strongly dislike chat are not practiced or comfortable writers in general, and, after some training (if such folks are willing, I've had them take documentation, presentation-building, or RFC/RFO writing tasks as practice), they tend to be much, much happier about (and better at!) textual communication.
I consider ITIL, Kanban and all other managerial erotic bureaucracy harmful. If I need to document something said by voice I have only to push a button on my deskphone and an mp3 of the conversation with reasonably good audio quality will be created and sent to my mail when I end the call, also I use to have a small portable recorder with a good mic in my pocket, while it's not legally valid it I do not tell others that I'm recording it turn out really useful in few cases.
Fortunately I'm in EU (France) so I have still a bit of choice (less and less, unfortunately) to avoid certain bureaucratic companies too much common in modern English world (and I evidence modern because when UK was a real power and USA was a really strong economy they do NOT have such modern mania or at least it was far far far less developed).
What I simply measure is how much I can do with single tasking concentrated operation vs how much I can do and how bad the resulting quality is when I multitask. The result are simple: the more I try to be quick the more error and time I waste, the more I go calm the best result I achieve.
And while I do not have real consistent data I can see result of a certain kind of work paradigm everyday in any aspect of our society.
