AFP Entrance study materials
AFP Entrance study resources

The Roth Conversion Before the Pension Starts

AFP Entrance · Put your knowledge to work in context.

All case studies

The Roth Conversion Before the Pension Starts

Devon, 58, is retiring from a state agency in two years. He will receive a defined benefit pension of $52,000/year starting at 60, with a 2% annual COLA. He also has $640,000 in a traditional 403(b) and $85,000 in a Roth IRA. His wife continues working and earns $60,000. Between retirement at 60 and Social Security at 70, their only income will be the pension and her wages. Devon's advisor proposes converting $80,000/year from the 403(b) to a Roth IRA for the five years before Social Security begins, paying the tax from the 403(b) itself. Their joint taxable income is currently near the top of the 12% bracket, and the 22% bracket begins at $96,950 (MFJ) and the 24% bracket at $206,700.

1. Is the five-year conversion window a sound strategy? Identify the tax-rate arbitrage and the risks.

Read the worked answer

The logic is classic bracket-filling: convert in low-income years so future required minimum distributions and taxable Social Security are taxed at lower rates. The window between pension start (60) and Social Security (70) is not actually low-income - the $52,000 pension plus $60,000 wages already put them near $112,000, inside the 22% bracket. So conversions will be taxed at 22–24%, not 12%.

The arbitrage only works if the conversion rate is lower than the future withdrawal rate. With a $52,000 COLA pension plus Social Security at 70 plus RMDs on a large 403(b), Devon's future marginal rate could easily be 22–24% as well - meaning little or no arbitrage, and possibly a worse outcome if he pays 24% now to avoid 22% later.

Risks: (1) paying tax from the 403(b) reduces the converted amount and may trigger an early withdrawal penalty if under 59½ - Devon is 58, so a 10% penalty could apply to the portion withheld for taxes; (2) conversions are irreversible under current law, so a market drop after conversion locks in tax on a higher value; (3) higher income can increase Medicare IRMAA premiums two years later and affect Social Security taxation; (4) the surviving spouse's filing status changes to single, raising future brackets - a reason conversions can still help.

2. What would you change about the proposal, and how would you present the trade-off?

Read the worked answer

Changes:

  1. Pay the tax from taxable/Roth assets, not the 403(b). Devon is 58; withholding from the conversion is a distribution that can trigger the 10% early-withdrawal penalty and shrinks the amount working tax-free.
  2. Right-size the conversion to fill the 22% bracket, not a fixed $80,000. Model the actual taxable income each year and convert only up to the top of the target bracket, watching IRMAA thresholds.
  3. Consider waiting until her wages stop - the truly low-income years may be after she retires but before Social Security, not the years she is still working.
  4. Partial, staged conversions rather than a five-year commitment, so a market decline doesn't lock in tax on inflated values.

Presentation: Frame it as a marginal-rate comparison - 'We convert only where the rate today is below the rate you or your survivor will pay later.' Show a projection of RMDs at 75 and the survivor's single-filer brackets. Be explicit that conversions are a tax-rate bet, not a guaranteed win, and that the biggest driver is the widow(er)'s bracket jump.