The management of the present Social Security System seems to have been operating in a panic mode for the past 10 years. Since 1994 the wage cap, which is officially called the “Contribution and Benefit Base”, has been increased from 60.6k to 90k for 2005. The increases are computed automatically using an algorithm that increases next years wage cap in the same proportion as last years increase in the national “Average Wage Index” another official reference value, but does not allow the wage cap to be decreased. The "Average Wage Index” is computed each year by taking the total or aggregate amount of wages subject to payroll tax that year and dividing by the total number of workers taxed to obtain the average wage per worker. This average wage is compared to the previous year’s value to compute the percentage change. Then the percentage change is used to change the Index to the same extent.
From 1999 to 2003 the average wage increased from 29.2k to 32.7k and it seems reasonable to conclude that this increase was in part due to the rising wage cap. It looks as though the algorithm was designed to ratchet-up the wage cap so as to capture an increasing amount of payroll tax money from the workers earning higher wages or salaries. It has a positive feedback property in that if an increased wage cap is followed by an increase in the “Average Wage Index”, then that will cause the wage cap to be increased still more (minimum increase is 0.3k). Some in Congress are not satisfied with the automatic ratcheting-up of the wage cap and are calling for a very large increase in the wage cap or removing the cap altogether.
The problem here is that it is just the workers making the higher wages or salaries that are paying additional payroll tax and they will receive little or no increase in retirement income for the additional taxes paid. Even worse, these are the workers that are most likely to face needs-testing at retirement time when some could be denied benefits if it is judged that they don’t need them.
Another problem is that for the past 7 years only 3 dollars of every 4 collected in payroll tax have been used to pay benefits. The extra dollar has gone to the surplus which over this same period has averaged about 147 billion a year, going from 107 billion in ’98 to 156 billion in ’04. Whenever the payroll tax revenue gets to be low relative to the cost of paying benefits , Congress has to raise the payroll tax rate and they raise it high enough so there is an excess amount of revenue provided. That way, they don't have to keep adjusting the payroll tax rate every year. The value of the surplus each year is added to the recorded assets of the Trust Fund but the cash is in effect transferred to the General Revenue Fund which consists largely of money obtained from corporate and individual income taxes. Consequently, the surplus each year is spent on non-retirement programs or budget items. The more serious aspect of this is that the surplus is frequently looked upon by Members of Congress as free money that can be spent on new or special programs or projects. This attitude is enhanced by the general belief that taking money out of the General Revenue Fund in order to refund surplus payroll tax would require raising corporate or individual income taxes and people don’t like paying higher taxes.
The Trust Fund has almost quadrupled in asset value in the past 10 years going from 436 billion in ’94 to 1687 billion in ’04. It is estimated that the Trust Fund will have a total value of roughly 4000 billion in 2017 when the annual surplus will have decreased from 156 billion to zero, i.e. assuming payroll taxes and benefits have not been significantly changed by then, which is very unlikely. The 2005 Annual Report of the Trustees of the OASDI Trust Fund states that, assuming no changes are made in the payroll tax rate or scheduled benefits, that it will be necessary to start using Trust Fund assets in 2017 and that the Trust Fund will be empty by 2041. At this point, the present payroll tax rate would only be sufficient to pay 74% of scheduled benefits and that this percentage would decline to 68% by the year 2079.
The Cato Institute report “Social Security Choice” states that for 2003 there were 3.4 workers per retiree and that the average monthly benefit was $888 per month or $10,656 per year, which is roughly the annual wage of a worker who works 40 hours a week at the minimum wage of $5.15 per hour. This is also slightly more than the poverty threshold of $9060 for a single person at least 65 years old and slightly less than the poverty threshold of $11,418 for a couple where the householder is at least 65. Using the $10,656 figure and the average 2003 wage per worker of 32.7k per year (an SSA figure cited above), it can be readily computed that the 3.4 workers can provide about 13.8k in payroll tax for each retiree. Also it can be seen that the average yearly benefit is about 77% of the 13.8k available and that this agrees well with the previous note that only $3 of every $4 in payroll tax collected goes to pay benefits.
Cato’s report also states that in 2030 there will only be 2.1 workers per retiree. Now the average benefit of $10,656 in 2003 is roughly 33% of the average wage per worker of 32.7k in 2003. Then if the SSA wants to maintain the ratio of average benefit to average wage to be at least .33 as the average wage increases due to inflation, it must keep the product of N and R at least as great as .33, where N is the number of workers per retiree and R is the total payroll tax rate. This is because the triple product NRA, where A denotes the average wage, is the maximum amount of payroll tax that is available to pay benefits on a per retiree basis. If R is maintained at 12.4% , then N must be greater than 2.66 or roughly 2.7 workers per retiree. It seems likely that the value of 2.7 will be attained around 2017 when the surplus is predicted to disappear. At that time it will be necessary either to dip into the Trust Fund or to raise the payroll tax rate, but using the Trust Fund could only be a temporary aid. If Congress were to raise the rate to 16% it would not be necessary to dip into the trust fund until at least 2030 and the number of workers per retiree might not go much lower than 2.1. Consequently, Congress could choose to raise the payroll tax rate to 16.4% , hoping to completely avoid using the Trust Fund to partially pay retiree benefits, but that would be a heavy burden to put on the workers, their employers, and the self-employed.
It is of interest to note that the Baby Boomers, born in the years 1946 to 1964 and which totaled about 76 million at birth, are currently in the workforce and will be retiring in the years 2011 to 2029. Consequently during this retirement period there will be a significantly above-normal decrease in the number of workers per year due to retirement and a corresponding above-normal increase in the number of additional retirees per year. It appears that the retirement of the Baby Boomers will precipitate the disappearance of the annual surplus in 2017 and the reduction of the number of workers per retiree to 2.1 in 2030, i.e. cause them to occur sooner than they otherwise would. This precipitation is like an early warning signal that our Social Security retirement system is failing us. It should not be construed that the Baby Boomers are in any way responsible for the impending failure.
Since it is expected that benefits will have to be cut under the current system, especially for the higher paid workers, tax reform could be used to help these workers. Congress could enact reform so that money paid in payroll tax would not be taxed as income, and even allow workers to put as much money into a tax-free Individual Retirement Account as they pay in payroll tax. That is, the money going into the IRA would only be taxed when it is taken out.