National Numeracy Network > NNN Columns > Dependency Ratios: Social Security and the "Celtic Tiger"

Dependency Ratios: Social Security and the "Celtic Tiger"

This material is replicated on a number of sites as part of the SERC Pedagogic Service Project


Dr. Eric Gaze
Director of the QR Program
Bowdoin College

Numb: In 2005 there were 62 million people under the age of 15 and there were 36 million people age 65 or over in the US.
Number: In 2005 the number of people in the US under the age of 15 (62 million) combined with those age 65 or over (36 million) represented 33% of the total US population.
Num-best: In 2005 the number of people in the US under the age of 15 (62 million) combined with those age 65 or over (36 million) was 33% of the total population; the share of the population in these age groups is expected to grow to 41% by 2050. In contrast, 49% of Niger's population in 2005 were "dependents," and the share of dependents is expected to drop to 37% by 2050.

The country of Ireland has undergone an astonishing turnaround in economic fortunes over the last quarter century, with the Emerald Isle now being compared to none other than Silicon Valley. As the "Celtic Tiger's" economic star has risen, the outlook for America's retirees and their federally-funded pension plan, Social Security, has waned. Certainly a case of correlation, not causation, you cry! While leprechauns are obviously not stealing coins from some pot of gold the United States has tucked away for its dependents, you can imagine all sorts of more viable social or economic explanations being proffered for each of these seemingly disparate situations. There is, however, a not so obvious connection between Ireland's boom and Social Security's thud: birth control. The United States had a boom of its own back in the Fifties, and now those boomers are heading into the golden years of retirement ready to collect all the cash they have been so dutifully sending to Uncle Sam to hold for them. Ireland, on the other hand, lifted restrictions on contraception in 1979 and is now reaping the benefits of all that protection. From latex to silicon in 25 years!

Sounds like something out of Freakonomics I know, but this idea actually comes from an excellent article, "The Risk Pool," by Malcolm Gladwell in the August 28, 2006 New Yorker. Pension plans, including Social Security, seem to be a no-brainer in terms of caring for retirees and their families. Employees pay a mandatory percentage of their salary into a pool matched by the company, which is invested by savvy professionals so that it compounds into a nice nest egg for the employee upon retirement. Excess gain from the pool is even used to cover disability claims and costs. So why are pension plans being phased out (except for the public sector) in favor of the mysterious sounding 401k plans?

The answer lies in dependency ratios, which compare the non-working age people (dependents) to those of working age in any given population; whether it be a company or a country. It might seem as though the number of dependents shouldn't matter since each employee is only collecting the money he or she deposited. Most pension plans, however, suffer from a fatal flaw: depending on how the plan was established, it owes money immediately to existing dependents. Known as a "pay-as-you-go" system, today's workers pay benefits for current dependents; and back in the 1950's the unions ensured the plans were very generous to existing employees and retirees who had paid in little to the pool.

In the case of Social Security, the Old Age and Survivors Trust Fund was well established in 1936, and had a balance of $267 million at the end of 1937 (US Fiscal Budget 2008). For twenty years this fund ran a surplus and the balance grew until 1958 when it paid out $8,041 million but only took in $7,825 million for a deficit of $281 million. Over the next twenty years the fund struggled to meet its growing obligations, with money coming in just barely covering money going out. The balance stopped growing and remained flat so that by 1980, the fund had an end of year balance of $24,578 million after paying out $100,615 million! In addition to the Old Age and Survivors Fund, three more funds were established: Disability in 1957, Hospital Insurance in 1967, and Supplementary Medical Insurance in 1967. By the 1990's all four of these funds were finally growing again, economically sound but demographically running out of time. The dependents were growing faster than the workers and the cost of care was skyrocketing. Thus the dependency ratio quantifies, in some sense, the "drag" on the financial well-being of that population due to supporting the children and retirees.

An example at this point will illustrate the devastating effect of dependents as captured in the dependency ratio:

General Motors' Dependency Ratios (The Risk Pool)
YearDependents:Employees
19621:11.6
20053.2:1

Back in 1962 each of those 11.6 workers was chipping in to cover the retirement of that lone pensioner, but 43 years later you have one exhausted employee trying to keep 3.2 retirees afloat. To make matters worse those 3.2 retirees are living a lot longer and paying a lot more for health care in 2005. There was just not enough money coming in, including interest gained on invested funds, to keep up with the massive outflow of cash for the pensions. I have heard that for each car GM sells, $1500 immediately goes to cover health care costs of employees. Now imagine a car company in a country with universal health care. That company can reduce its price by $1500 and still make the same profit as GM! What happens when a company can't keep up with its debt obligations? It goes bankrupt and those pensioners are left without a dime.

The stark calculus of assessing a company's pension burden with a dependency ratio may seem too simplistic and cold-hearted to apply to a country's well being. Surely the old and young provide many intangible benefits that cannot be so crudely quantified? You, of course, can be the judge. In 1970 the fertility rate in Ireland, meaning the average number of children a woman could expect to have in her lifetime, was 3.9 (The Risk Pool); in 2005 it had basically been cut in half to 2 children (2005 World Population Data Sheet). As a side note, the highest fertility rate in 2005 goes to Niger at 8 children with Afghanistan close behind at 6.8 children per woman. In the 1960's Ireland's dependency ratio was 1 dependent for every 1.4 worker bees, while in 2005 the ratio had changed to 1 dependent for every 2.2 workers (The Risk Pool).

The dependency ratio as defined above is the ratio of non-working age people to those of working age which begs the question: What is the working age? The way we define this, as Joel Best in Damned Lies and Statistics so admirably admonishes us, is critically important to the social construction of this statistic. The U.S. Census Bureau uses the age group 18-64 years old as the working age for their dependency ratios. Every other source I can find defines working age as 15-64 years old, so we will use this latter definition.

Dependency Ratio =[Under 15 plus 65 and over]:[15-64 year olds]
Dependents:Workers

Next we must decide how to represent our ratios. Malcolm Gladwell chose to scale the number of dependents to 10, giving us the dependency ratio 10 :14 for Ireland in the 1960's; which I then chose to rescale to 1 : 1.4. Most websites would represent this ratio as a decimal, 0.714, i.e. with the second number scaled to one: 0.714 : 1; while the US Census Bureau would have us multiply by 100 to get 71.4. The Census Bureau is, of course, using the unfortunately familiar grade school technique for "converting" a decimal to a percentage, but multiplying a decimal by 100 to make it a percentage makes no sense, since 0.714 = 71.4/100 = 71.4 per 100 = 71.4 per cent = 71.4%. Interpreting this ratio as a percentage is very tricky since we are comparing two parts of the whole population, leading to the sentence: the number of dependents is 71.4% of the number of workers. It is much better to say there 71.4 dependents for every 100 workers, or 714 per 1000. I calculated the dependency ratios in the following table using the International Data Base from the US Census Bureau to 3 decimal places, cross checked the 2005 ratios with the 2005 World Population Data Sheet, and the 2005/2025 ratios with figures from the Department of Labor:

Country200520252050
Ireland0.4810.5440.748
United States0.4930.6150.691
China0.4290.4730.613
India0.6060.5040.519
Niger0.9720.8690.586

You can read these numbers per 1000, so that in 2005 Ireland had a very low 481 dependents per 1000 workers, which will jump back up to their 1960's number of 748 dependents per 1000 workers by 2050. Note how Niger's ratio, like most of Africa, is much higher than the more developed nations and certainly a factor in these third world countries' failure to accelerate their economies. The developed nations are all aging however, which accounts for their rising dependency ratios. According to the Organization for Economic and Cooperative Development (OECD), in 2005 roughly 12% of the population of most developed nations was 65 and over. This percentage will basically double by 2050 so that one-quarter of these countries' populations will be senior citizens. This may be the death knell for Social Security and why so many people are predicting the demise of this program. We just won't be able to support all of these dependents without raising the tax on workers to unsustainable levels. Niger and India will not have this to worry about so maybe one of them will be the next Celtic Tiger!

"Sapere Aude!"


The author welcomes any suggestions of ratios for use in this column. If your ratio is chosen (first suggestion received for that ratio) you will be sent a complimentary Alfred University t-shirt!


« Planck's Constant: Motion and Rest