Monday, April 9, 2012

Social Security: Thinking about Pay As You Go

Robert J. Samuelson has a column in today's Washington Post questioning whether or not Social Security today is the program that Roosevelt wanted. I don't know the history well enough to have an opinion, but his statement that "Roosevelt rejected Social Security as a “pay-as-you-go” system that channeled the taxes of today’s workers to pay today’s retirees. That, he believed, would saddle future generations with huge debts — or higher taxes — as the number of retirees expanded," resulted me thinking about pay as you go systems. My main thought was, for an economy as large as the United States, what could a Social Security system possibly be other than pay as go?

As Samuelson points out, "in 1960, there were five workers per recipient; today, there are three, and by 2025 the ratio will approach two."

This fact remains whether or not Social Security was pay as you go. What difference would it make if it had been saved? We could perhaps have bundled dollar bills paid in into giant bales and stored them in a vault until needed, but this would do nothing but drive inflation when they were taken out. We can do this anyway through monetary mechanisms to drive production and reduce debts without the absurdity of bailing dollar bills.

Alternately, this money could have been invested in capital markets, but capital depreciates and, in aggregate, has diminishing marginal returns. What good would this additional capital have done, it would be depreciated by the time the labor crunch comes. This has the same problem as pay as you go, the labor supply ratio is still less beneficial and the capital is being invested at the wrong time for the wrong labor force conditions. Trying to drive investment beyond its efficient level will result in little more than bridges to nowhere and stock buy-backs. What good would this do?

A more sensible alternative would be foreign exchange reserves. But even here, most countries have demographic problems similar to the US. Accumulating these reserves would involve considerable risk, the stable economies are facing the same labor force problems we are so trade among a group of countries with declining labor force to population ratios doesn't achieve much. Currency reserves in emerging markets might help, but this is gambling with our savings. Go back and read about some of the greatest hopes of the 1950s or 60s, few of these are grand success stories today. This approach would achieve something, but not that much and would have faced substantial political opposition.

Another, rather absurd, approach would be to store commodities to drive prices down and raise output when demographics turned unfavorable. This would have the impact of slowing growth when demographics are favorable and raising it when demographics become unfavorable. However, there is no way to be certain what future resource demand will be, technological shifts may make some resources lose value. Also, it is not certain that current investment would not have a higher payoff than stored commodities.

Ultimately, while this column has some good things to say about the optics, people do feel their Social Security contributions are safely "saved," it retains the flaw that while an individual or small country can defer consumption an economy the size of the US really can't. For the US as a whole we can't defer consumption till later by saving, we can produce whatever current factors of production can produce. We can attempt to invest to increase future production, but this doesn't impact future dependency ratios, the system is still ultimately pay as you go. At an aggregate level all Social Security funds have to be currently spent, there is simply no way to save them. All we can do is choose whether we will spend them to increase our capital stock, increase the quality of our labor force, reduce current obligations to other creditors, or squander the money on useless projects.

Given that the increasing dependency ratio is the major factor here, the most sensible outcome would have been to focus on labor force reproduction and skill levels. However, much of this gets labeled as consumption, despite strong data indicating that spending on healthcare, income security, and early childhood programs are strongly associated with future earnings and income mobility. Unfortunately for us, little thought is given in the US to the merits of the development state, we remain bound to dated 19th century notions of the economy that exaggerate the role of capital while ignoring the tendency to overinvestment when the labor force has failed to keep up with capital spending. The US isn't a household, programs like Social Security can't work for the US as a whole the way a savings account works for a household. We need to increase the factors of production and in the US land is basically fixed, we have an abundance of capital, but our labor force, particularly the future part of it with low income parents, isn't in the best shape. Other countries have done better, the US has been declining in the rankings of labor force quality, such as educational attainment, as our social programs have failed to expand at international rates (50 years ago the US had as strong, or stronger, social programs than other countries). We don't have much time left and thinking about Social Security as saving isn't going to help us. We have to focus on labor and the dependency ratio whose increase is the ultimate source of the problem. Unfortunately, I don't see that happening.


  1. The problem with pay-as-you-go is the change in the dependency ration. And the reason that has changed appears to be that life expectancies have increased, but retirement ages have not.

    A century ago, it was routine for someone to keep working far longer. If you were still alive at 90, you were likely to be working at something still. Maybe not the career you followed earlier, but some kind of job was normal. You "retired" only for the last year or two of life, when you were physically beyond doing much of anything.

    When Social Security was started, life expectancies were just a handful of years beyond the eligibility age. But today....

  2. I'd agree that life expectancy changes mean that there need to be some major changes to Social Security, but I don't see this as linked to pay as you go. If we had delayed benefits until they were fully funded I don't see that this would have made a difference, which is what I normally see mentioned in relation to pay as you go.

    Regarding life expectancy, Social Security reform did raise the retirement age, I'm facing a retirement age of 67 instead of 65. Further tweaking is likely necessary, but the discrepancy in age 65 life expectancy by income means that there are some equity concerns regarding simply raising the age. I mentioned in an earlier post that I think capping the income received based on age with caps ending around age 70 would be the best compromise.

    I'd also dispute the idea that a century ago people worked longer. Poverty rates among the elderly were around 50% back then (and as high as 33% in 1960 before programs expanded). Also, most (though not all) of those people working till 90 were probably in sinecure positions, which are unlikely to exist in a work environment as competitive as today's. Today I'm sure that more could be done to encourage the elderly that can to work longer, but there are a lot of people past 65 that simply aren't competitive employees anymore, particularly those that have held low wage occupations their whole lives.

    Of course, all of this was expected when Social Security was originally instituted. There was no agreement over how to adjust benefits to inflation or how to adjust taxes and other aspects of the program. So they decided to muddle through. It's been necessary to reform the program at certain times, and this remains necessary. That's completely normal for a program that lasts for decades and I don't see the pay as you go aspect having much impact on anything. It just makes an explicit link between benefits and the growth of the US economy, which is inevitable, whether or not politicians attempt to hide it, with a program that is significant as a portion of world GDP.

  3. What I was trying to get to was, if the age of retirement had gone up anywhere near as much as the life expectance at age 65, the dependency ratio would not have changes as much as it has. That is, there would be more workers available to pay into "pay as you go," and fewer taking money out. Which would make it more viable.

    The issue of poverty among the elderly is certainly something that needed to be addressed. But I'm not sure that Social Security with a static retirement age is the best approach to that. Also, in passing, I note that today's general anti-poverty programs were unknown when Social Security was created.

    I'm not sure I totally subscribe to your point about people being no longer able to work past 65. For people in physically demanding occupations, sure, but that's a declining demographic. Increasing numbers of people are in jobs where the physical exertion required is picking up a telephone or typing on a keyboard. I see no reason why that is beyond someone who is 70 or 80.

    Perhaps we need to muddle thru to a way to differentiate between what careers can and cannot be extended. Or perhaps we need to think about how to help people to shift careers when they cna no longer handle the physical demands of their first. Admittedly, it will take a significant cultural shift for some to accept that they might need to learn something new after about age 20. But that is a problem of willingness, I believe, mroe than one of ability.