Suppose that you have a client. The client, after thorough consultation with you, wants you to take some action. That action is in the best interest of the client, is legal, and is ethical.
You do it, right?
Now suppose that the action is not in the client’s best interest, but is legal and ethical.
Again, you do it. It’s not your call.
Now suppose that the action is in the client’s best interest and is ethical but is illegal.
You don’t do it.
Okay. Now it’s in the client’s best interest and legal, but is unethical.
Do you do it? No, of course not. Your ethics trump even the client’s interests.
But wait. Suppose that the only person who might be harmed by the ethical violation is the client, and that you’ve advised the client of this possible harm, and the client still insists that you take the action.
What do you do?
A ridiculous hypothetical? Unimaginable? Highly speculative? Nope. An everyday situation in federal criminal defense.
There are a few possible answers: advise the client to waive the 2255 (ethics be damned); advise the client to refuse the plea (client be damned); refer the client to conflict counsel (to advise the client independently on whether to waive the 2255); or move to withdraw (because there’s an irreconcilable conflict between your interest in maintaining your ethics and your client’s interest in the action).
None of them is a great answer. “Refer the client to conflict counsel” is least bad, but most criminal defendants in federal court don’t have resources to hire another lawyer to advise them on whether to waive the 2255. (Though asking the court for resources for such counsel, either under the CJA or in an Ake motion, might bring the point—that the government should not be asking for such a waiver in the first place—home to a federal judge better than anything else.)