2003 pre-EGTRRA 415 $ Limit

[David MacLennan]

I think I have this right, but for confirmation I was wondering if anyone else has computed this.

For DB plans not reflecting EGTRRA in 2003 valuation, 2003 415 $ limit:

($90,000) * (sum of 2002 3rd Quarter CPI-U) / (sum of 1986 4th Qtr CPI-U)

= ($90,000) * 541.8 / 331.2

= $147,228

Rounded to next lower $5,000 mulitple,

= $145,000

[Guest Doug Goelz]

I get the same result of $145,000.

However, for a really insignificant note, Jim Holland supposedly said in a 2/4/93 letter that the IRS uses a value of 331.3 (vs. your 331.2).

[David MacLennan]

The 0.1 difference is probably from the CPI-U table I have access to, which was rounded to 0.1. I think the 331.3 is rounded after summing the CPI-U quarterly numbers with more decimals.

[MGB]

There is no CPI figure ever produced with more decimal digits. The current reporting of the 3 months from 1986 is obviously 331.2, not 331.3. Although I am really stretching my memory here and trying to use some brain cells long since lost, this may have been one of the periods that were later adjusted it to reflect an error in methodology at the BLS in compiling the index. This has happened a number of times in the past few decades. So, the IRS may have written down the original 331.3 and never changed it when BLS later changed it. (Alternatively, the IRS may have just gotten this wrong and somewhere wrote down 331.3 in error and have used it ever since.)

I have a fax from the IRS from 1998 on their exact methodology. It does refer to the 331.3 (without any explanation as to why it is different than BLS's current reporting, but does state it is the sum of the reported amounts from BLS, which are always to one decimal place). There is one more rounding in the algorithm that is done that is not reflected in the prior posting:

541.8/331.3 = 1.63537579....

This step is rounded to four decimals (1.6354) before multiplying by the statute amount. Note that there are other statute amounts other than the 90,000 that are multiplied by this same rounded factor.

So, it is really 1.6354*90000= 147186 (not 147184, which is 90000*541.8/331.3), and then rounded down to 145000.

In the example from the IRS, they gave the result of this multiplication as a whole number, but did not state that any rounding occurred. I think it is actually an exact calculation, becuase they go on to state that in order to step up to the next 5,000, it would have to be at least exactly equal to that 5,000 step.

[MGB]

According to the following page, the US indices for the time periods involved have never been changed (although other numbers have changed a few times in the past):

http://www.bls.gov/cpi/cpirev01.htm

I guess that leaves my "choice 2" as the probable situation. I.e., someone in the IRS incorrectly recorded the 331.3 figure.