Whose numbers are right?

[National Dialogue Moderator <moderator>]

Throughout our discussions, we have seem various references to 
different studies, and it is clear that they reach different 
conclusions in their analysis.  I want also to make a distinction 
between the projections by the social security actuaries and outside 
studies by experts.

Official projections:  A review of the work of the social security 
system actuaries will show that they do not produce only one cost 
estimate each year.  Rather, they produce three, to provide a range of 
results. The size of the expected actuarial deficit is very different 
in the three estimates, so that the estimates drive our notion of 
whether there is a problem.  I would like to encourage those who want 
to understand the numbers to focus on the range and the uncertainty as 
well as the specific numbers.

The long term costs and benefits under any alternative depend not on 
the estimates made in advance but on the experience under the system.  
Some of the things which affect the actual costs include:

  1. How many people are working, what they earn, and how many are 
     unemployed
  2. The size of the working age population which depends on rates of 
     birth, death, disability, immigration, ages of retirement, etc.
  3. When people retire
  4. Rates of wage growth
  5. Rates of inflation
  6. When people die

There are many more variables.  A great deal of work is done to 
produce estimates which are as accurate as possible, but conditions 
change, and what appear to be small changes in some of these variables 
can have a big impact over a long period.  While there is not 
unanimous support for the official projections, I would like to point 
out that the social security system has a very qualified actuarial 
staff, and that in additional, there have been a number of technical 
panels of experts who have repeated reviewed the assumptions and 
methodologies used in these calculations.  

Outside studies:  Many outside studies have been undertaken by 
experts, with different qualifications and agendas. The results reach 
very different conclusions.  The results are very sensitive to the 
assumptions made about investment earnings, when people will die, cost 
of running the system, etc.  Many of these studies show single answers 
rather than a range of results.  Before accepting the results, it is 
important to understand the assumptions and methodology used in the 
calculations.  Some of the questions to ask include:

  1. What rates of return have been assumed?
  2. Have increases in wage rates been projected on the basis of an 
     economic scenario which is compatible with the investment returns?
  3. What is the assumption with regard to earnings sharing?
  4. What administrative costs have been assumed, and how has it been 
     determined that they are achievable?
  5. If it has been assumed that death and disability coverage will be 
     separately insured, has adequate provision been made for the cost of 
     this coverage including coverage for those who are not healthy?
  6. How are transition costs provided for?

For more information on the actuarial issues involved in understanding 
social security costs, contact the American Academy of Actuaries 
(www.actuary.org).


Anna Rappaport