beginning IRA/40k distributions in 2010
I turned 70 1/2 earlier in 2010. I realize I can begin taking required minimum distributions (approximately $26,000) from my IRAs and 401k either by 12/31/10 or by 4/1/11. In addition I must take a distribution (approximately $3,260) from an inherited non-spouse IRA by 12/31/2010. I’m in the 15% tax bracket and on social security. A good portion of my social security in not now taxed. I think I may be better off (tax-wise) to take the $3,260 distribution in 2010 (thereby keeping some of my social security tax free, and take the $26,000 in April 2011. I do realize that I have to take another distribution (for 2011) by December 31, 2011, which will mean, of course, all my social security will be taxed. Am I correct? Are there other tax implications that I thinking about?
Permalink Submitted by mk foss on Thu, 2010-09-09 02:35
If you defer your first RMD until 2011 – you save the tax on Social Security for 2010 as you indicate. In 2011, you will have two RMDs – if it’s 26K for the 2010 RMD it’s likely to be more for 2011 – because hopefully the value will have gone up plus you have a larger percentage by being a year older. If Congress does nothing – tax rates will go up in 2011 and the 10% bracket will be eliminated. Also having such a large income in 2011 could put you into the territory where your Medicare premiums are increased for a year. The increase occurs 2 years after the big spurt in income.
Another factor is that if you do not take the RMD in 2010, you’re precluded from making a Roth conversion in 2010. The Roth conversion can only occur after you’ve taken your RMD for the year even though the law allows a deferral to the next year. An advantage of the Roth conversion is that you may be able to reduce future RMDs by prepaying tax (either in 2010 or using the 2011/2012 default).
You’ll be taking RMDs from your IRA, your 401k and the inherited IRA for every year going forward. I don’t know if saving the tax on social security for just one year (especially if it affects your Medicare premiums) is worth it. You can try waiting to see if Congress does any thing about tax rates (probably nothing will occur until after the election) but then you may need to rush to get that RMD into 2010 if that’s your choice.
Good luck.
Permalink Submitted by Alan Spross on Thu, 2010-09-09 02:55
This is a confusing issue. As your SS income is being added to AGI @ 50% and then 85%, the net affect on your marginal tax rate is a 50% boost or an 85% boost. For example, if you are in the 15% bracket, while SS income is being added due to increases in other income, your actual marginal rate is 1.5 or 1.85 times 15%. When 85% is being incorporated, your total rate is then 27.75%, actually higher than your next basic bracket of 25%. So the rate drops right after all your SS is included at 85%. Therefore, if you fall in the right tier of your bracket, you can save taxes by staggering your Roth conversion in uneven amounts in different years. This also applies to your RMDs in your first distribution year. But this is a mind boggling exercise, even assuming the only income you have control of is your conversion or your RMD for this one year.
You might pull up a SS calculator off the web, and just enter all your other income and then experiment with that RMD in various combinations. You are able to take none of any part of that first year RMD in the first year. Then use the same calculator to estimate 2011 taxes when you include income that was not taken in 2010. I think you are reasonably safe in using the same tax rates and brackets at this point since It is unlikely that rates for incomes under 200/250 will rise in 2011. But keep tabs on that unpredictable Congress as well. The estate tax fiasco proves they are capable of anything.
I am busy trying to do the same thing only with varying the start date of SS benefits instead of RMDs to determine how much to convert and when to stop generating delayed SS retirement credits. It’s a mind boggling exercise and I have the constant feeling I am overlooking some variable.