Yglesias' argument is better than Jack's. In fact, his reasoning is similar to the reasoning for attorney-client, physician-patient, and priest-penitent privileges against testifying in court. But I'm not quite persuaded by it, for several reasons:
- As Yglesias notes, in-person and over-the-phone conversations aren't subject to FOIA. Why isn't that accommodation sufficient for frank discussions? It seems to me that at this point, we're balancing two interests: that of the government employees in having frank discussions via email (or Whatsapp or what-have-you)--not their interest in having frank discussions at all--and the public interest in government transparency. Isn't it obvious which should prevail?
- Email isn't exactly like letters or memos, but neither is it exactly like a personal chat. Maybe it shouldn't automatically be treated like the former, but that doesn't come close to showing that it should automatically be treated like the latter. Unless Yglesias has some better system of sorting emails into disclosable versus nondisclosable than presumed-non-disclosure, I think we should simply err on the side of disclosure.
- It's not clear to what extent fear of FOIA leads to a different outcome than a general fear of leaks.
- We have good reason to want records from executive agencies regarding inputs, as well as outputs. We ought to know why they're doing what they're doing, and not merely know what they're doing.
- At bottom, the problem is frank conversations that are (or could become) politically embarrassing. That's not a good reason to exempt the conversations from disclosure--in fact, the whole point of FOIA was to remove that as an excuse for non-disclosure.
You later suggest that FOIA should include more exemptions. Depending on the exemptions, I might or might not agree with that, but it's still a much better suggestion than Yglesias' presumption against disclosure.