At the AICPA conference, and then later in a Journal of Financial Planning article, I read a really cool idea. I’m not sure it’ll still be possible after Congress gets done hashing out changes to the tax code, but I’ll note it here and we can circle back in January.
Here’s the trick: let’s say you’re thinking of converting between $10,000 and $20,000 of IRA money this year. You don’t know for sure how much room you’ll have in your tax bracket, but somewhere in that range.
In January 2018 you create three separate Roth accounts and you do three separate Roth conversions, each for $10,000. One is large cap domestic stocks, one is emerging foreign market stocks, and a third is for intermediate government bonds.
The next spring when you’re doing your taxes (or Fall, 2019, if you’re on extension) you check to see what each of those three $10,000 Roth conversions is worth now. Then you recharacterize back to a Traditional IRA which ever one did the worst. You keep the one that went up the most, and the one in the middle you only keep as much in a Roth as you had room for in your tax bracket (which is now determinable) and recharacterize the rest back to a Traditional IRA.
This allows you to convert to capture the most possible gains tax-free inside the Roth, leaving the losers and slowest growing back in the Traditional IRA.
This trick relies on our ability to recharacterize a Roth conversion back to a Traditional IRA up to the due date (including extensions) of the tax return. Rumor has it that this loophole is on the chopping block. Let’s stay tuned.