Policy
Simulation
Group

Pension Projection Models

Enhancements Introduced in Current 2/17/16 Version of PSG Models

  1. Rename GEMINI .sum output statistic earn50 to earnAA and add to GEMINI STATS table a new input parameter called earnAA_age. The baseline value of STATS.earnAA_age is 50, so the .sum output file is unchanged unless the value of STATS.earnAA_age is changed. This enhancement simply allows the output of earnings at ages other than fifty.

Enhancements Introduced in Earlier 2/5/16 Version of PSG Models

  1. Add three new GEMINI STATS table input parameters that control how average-final-earnings replacement rates are computed. The Social Security Advisory Board's 2015 Technical Panel on Assumptions and Methods recommended to the SSA Office of the Chief Actuary that the annual Trustees Report include replacement rates that are expressed as a percent of average final lifetime earnings. Pages 83-85 of the Technical Panel's final report contain their recommendations and some preliminary results. Recently CBO has adopted a version of the Panel's recommendation and presented replacement rates expressed as a percent of average final substantial lifetime earnings in the recent CBO's 2015 Long-Term Projections for Social Security: Additional Information report on pages 3-4 and 16. The three new GEMINI STATS table input parameters allow users of the PSG models to calculate a wide range of these kinds of average-final-earnings replacement rates. The three new parameters do not change any output statistics generated by prior versions of GEMINI.
         The first new GEMINI input parameter is STATS.eh_awi_pct, which determines how large annual earnings must be to count that year as a lifetime earnings year. CBO computes replacement rates only for individuals who have at least twenty years in which annual earnings are at least ten percent of that year's AWI. The value of this new parameter is 10 in the PSG benchmark runs, which is the same as used by CBO.
         The second new GEMINI input parameter is STATS.rr_aie_pct, which determines how large annual earnings must be to considered substantial, and therefore, be included in the average of final earnings, which is the denominator in these replacement rate statistics. CBO includes in the average only annual earnings whose real (age 62 dollars) value is at least fifty percent of the individual's AIME*12. The value of this new parameter is 50 in the PSG benchmark runs, which is the same as used by CBO.
         The third new GEMINI input parameter is STATS.rr_n_years, which determines how many substantial earnings years are included in the average. CBO includes five of the most recent (that is, final) substantial earnings years in the average. The value of this new parameter is 5 in the PSG benchmark runs, which is the same as used by CBO.
  2. Add to the GEMINI .sum output file three new statistics, all of which are related to average-final-earnings replacement rates (see the background discussion just above). These new statistics have been added to the end of each row in the .sum output file, so that existing .sum output statistics are unchanged in their numbering (and, of course, are unchanged in their value).
         The first new GEMINI .sum output statistic is (48) eh_yrs, which is the number of years annual covered earnings equals or exceeds STATS.eh_awi_pct of AWI in that year.
         The second new GEMINI .sum output statistic is (49) afsern, which is the average final substantial real covered earnings expressed in thousands of dollars for the year the individual is age 62. Substantial is defined by the STATS.rr_aie_pct input parameter and the number of final years included in the average is defined by the STATS.rr_n_years input parameter. This statistic is the denominator of the replacement rate.
         The third new GEMINI .sum output statistic is (50) aime62, which is AIME*12 expressed in thousands of age 62 dollars per year (rather than dollars per month). Notice that some individuals will a zero value for aime62 because they did not have an eligibility event at age 62 or they did but were ineligible for benefits at age 62. This statistic has been included in the output because it has been used by CBO and the SSAB 2015 Technical Panel to compute a hypothetical social security benefit that is the numerator of the replacement rate.
  3. Add to the PSG models distribution an AWK script that shows how to compute CBO-style replacement rates using the new GEMINI .sum output statistics. The script is named CBO-rr.awk and is located in the SSASIM\db directory; it computes replacement rates for a subsample of an RCS-mode cohort sample. Read the short script to understand what revisions need to be made in it in order to calculate replacement rates for individuals not born in 2004. The CBO-rr-540.rsf whose output is tabulated by this script has also been added to the PSG models distribution.

This page was last revised on February 17, 2016.