Yesterday, Howard explained why the Roth rollover loophole, set to open in 2010, is horrible policy. By pre-paying tax on accumulated IRAs, rich folks can effectively increase their tax-free savings by close to half. (My 2006 article explains this point.) And the rollover option is worth most to people who don’t actually need retirement accounts to finance their old age.
The policy doesn’t serve any valid goal, other than to further enrich the already wealthy. If Obama wants to raise taxes on high-income Americans, he could repeal Roth rollovers at virtually no economic cost (i.e., it wouldn’t affect work, saving, or investment).
As Howard noted, however, the obstacle is budget scoring. Roth conversions ostensibly improve public finances in the short run. And, under standard scoring rules, JCT and Treasury would have to score repeal of the rollover provision as a revenue loser, even though it would save taxpayers tens of billions of dollars over the long-term.
In fact, the rollover gambit is really borrowing—on very unfavorable terms to the government. I estimate the implicit interest rate to be 14 percent. That obviously makes no sense when the government can borrow long-term at around 4 percent.
If CBO and JCT had scored the rollover provision as borrowing rather than revenue, Congress wouldn’t have been enacted it. The revenues produced from the conversions would have been disregarded and the substantial and growing revenue losses would have been scored as a multi-billion dollar cost over the budget period.
Roth-type shenanigans could also be prevented by showing revenue estimates on a present value basis, at least under circumstances like this where the long run and short run implications of a policy are polar opposites. (This might be a useful way to look at some health reform proposals as well.)
Changing the scoring rules from cash accounting raises thorny issues, which I haven’t fully thought out. But if we want to get our finances on a sound footing, a good way to start would be to stop cooking the books.
