(November 2008) The U.S. population is aging: The ratio of elderly to the working-age population in the United States will roughly double over the next few decades, straining the finances of the U.S. Social Security system and other government programs.
During a PRB Discuss Online, Ron Lee, professor of demography and economics at the University of California-Berkeley, answered participants’ questions about the trends in mortality and aging in the United States, and the implications of population aging on government entitlement programs and the U.S. economy.
November 6, 2008 1 PM EST
Transcript of Questions and Answers
John Rohe: Hi Ron, Here is my article on the subject (pertaining to the US) as of 2003: http://www.thesocialcontract.com/pdf/thirteen-four/xiii-4-248.pdf. There are two ways to impact the dependency ratio: retirement age and immigration. Are the figures from 2003 (in my article) still accurate? At the time, Leon Bouvier assisted with the demographic computations.Thank you, John Rohe
Ron Lee: John—I read your article; thanks for sending it to me. The outlook for US demography has not changed much since 2003. However, there are a couple of points that struck me. You mention that the TFR had fallen to 1.7 in the mid-1970s, which is approx true; however, it subsequently rose to replacement level and has been bouncing around between 2.0 and 2.1 for many years. So this is a sharp contrast to the other industrial nations which generally have considerably lower fertility. Second, you refer to the UN study on the amount of immigration that would be required to maintain the old age dependency ratio at its 1995 level through 2050, and report that the UN calculates a total of 593 million would be needed, if retirement remains at 65. I myself did a calculation of the pace of immigration that would be required to make Soc Sec finances balance through 2075 (that is, make the summary actuarial balance equal roughly zero) and found 5 million per year, which over 75 years would be 375 million, a bit less than the UN number. However, the basic conclusion is the same as yours either way: that level of immigration is not an appealing policy option.
Tina Dutta: In this advanced era of living, where medicines and lifestyles have improved a lot with low birth and death rates, aging of population is an inevitable outcome in most of the developed countries. Of course, the age pyramid can not be reversed now immediately, but how best can we utilize our human resouces [so] that the elderly population can be considered as productive- financially and socially rather than unproductive consumers only?
Ron Lee: Tina—I agree that population aging is inevitable. I think that many of our institutions that are impacted by aging are rather inflexible, and were originally developed to be effective in a different demographic context. Social Security, Medicare, Institutional Medicaid, many private pension plans—all these get into fiscal trouble with longer life and population aging. I think we need to redesign these institutions so that they at least permit individuals to respond to their circumstances in ways they views appropriate. Social Security has gone a good distance in this direction, trying to present workers with a fair tradeoff between continuing to work longer and getting a higher pension when they do retire, or retiring earlier and receiving a lower pension. Personally, I don’t think policies should take for granted that the best solution is for people to retire later. I think we should let people choose, but set things up in such a way that their choice does not impose costs on others. In fact, I would like to see our institutions permit leisure (time not doing wage labor) to be taken throughout the life cycle as a person chooses, rather than being all bunched together at the end of life.
Barbara Haley: Even though the ratio of elderly to the working-age population in the United States will roughly double over the next few decades, the dependency ratio is going down. What alternative to the regressive payroll tax (that currently funds old age) should the US adopt, to take full advantage of this?
Ron Lee: Barbara—The Total Dependency Ratio is the ratio of the youth population plus the elderly population to the working age population. A more refined measure is the Support Ratio based on empirical age profiles of consumption and labor income, used to form the ratio of equivalent workers to equivalent consumers. This ratio declines at .2% per year from now through 2050 (analogous to a rise in the dependency ratio). This rate of decline is, I think, very slow, particularly in comparison to the much bigger changes in the finances of the Social Security system or Medicare. But by construction, it is considering not only these govt programs targeted to the elderly, but rather all govt programs and also the age patterns of private consumption. You raise a good point about the payroll tax being regressive. I agree. The original idea was that the progressivity of the benefit schedule would out-weigh the regressivity of the payroll tax, but because poor people die younger than rich people, it is not clear whether actual benefits are progressive. One could, of course, fund old age benefits out of the income tax, but I think most people would rather see a tighter relation between what benefits you receive and what contributions (payroll taxes) you pay. One way to do this is the so-called National Defined Contribution systems that are common in Europe, but these do nothing to make the system more progressive. That then requires a separate redistributive component, which some plans like the Swedish one have.
Richard Cincotta: Commentators on population aging regularly make statements like, “the U.S. population is aging much slower than either China’s or Japan’s.” Yet, only once (a US Census Bureau/IPC slide) have I seen a metric to assess such a rate. What metric would you recommend for comparing this phenomenon?
Ron Lee: I don’t recall the details, but the metric I have often seen is something like the number of years it took for the proportion 65+ to go from 5% to 10%. Those thresholds aren’t right, but I think that concept is the one most commonly used.
Cecily Westermann: Initially, Social Security Old Age benefits were available only to workers who contributed to the system. Later this benefit was extended to “non working spouses”. According to the Department of Labor, 73% of women with children between 6 and 18 (no child under six) are working, therefore earning their own benefits. If Social Security is truly expecting a shortfall, why can’t non-working spousal benefits be phased out of the program?
Ron Lee: As I recall, a person gets to choose whether to take a Soc Sec benefit based on their own earnings history or on that of their spouse (or someone to whom they were previously married for at least ten years), where they would be entitled to 50% of the spouses benefit level. I am not sure I have that exactly right, but that is the general idea. Female labor force participation is really quite high, as you note, but many of those women have had quite interrupted work histories and lots of part time work, while they took time out to raise children. The benefit they would qualify for based on their earnings histories would often be very low. And for those women who were out of the labor force for most of their lives, the benefit level based on their earnings might be zero (if they didn’t work for at least ten years). I would prefer other policies to deal with the projected shortfall in Social Security.
Emmanuel Amodu: What are the financial institutions doing about [the] lending rate to [retired] and aging people? Are there special packages for them? thanks
Ron Lee: There are reverse mortgages, but few older people use them. there are also special investment funds that automatically adjust the risk-rate of return tradeoff as people age. Probably there are many others, but this is not an area I know much about, despite the title of this discussion. However, I suspect that there is a lot of room for development of other special financial instruments for the elderly.
Rune Bakken: Some European countries are facing the same shift in numbers of elderly vs the still working population. Will an altered immigration policy in the West improve this ratio as an attempt to maintain socioeconomic stability, simultaneously alleviating the imbalance of young vs old in e.g. the Middle East and Africa?
Ron Lee: Immigration as a way to moderate the extent of population aging in the industrial nations is, of course, quite controversial. From a demographic point of view it certainly has some effect on the age distributions in the receiving industrial nations, and less so in the Third World sending nations because the that population is so much larger. However, the demographic effects on the receiving countries’ old age dependency ratio are surprisingly small, particularly in the longer run, since the immigrants grow old themselves. Also, the TFRs in the sending countries are often not much higher, and may be even lower, than in the receiving country (e.g. China and the US), so there will be less and less affect on fertility in the industrial nations. There is a fiscal gain per immigrant in the US, given the current composition of the immigrant stream, and that effect is large on a per immigrant basis but not very big on a national basis (see Chapter 7 in the New Americans, a Nat Acad Press report).
Kelvin Pollard: What remedies for the effects of population aging on entitlement programs have Europe and Japan tried? Can some of these efforts (in full or modified form) be applied in the United States?
Ron Lee: I should know more about this than I do. The main thing that comes to my mind is switching to National Defined Contribution (NDC) public pension programs in Sweden, Latvia, Italy, and to systems with similar features in Germany and France. These systems are designed to mimic Defined Contribution systems (like 401Ks in the US), in the sense that each individual has a (imaginary) fund based on the amount contributed in the past and a rate of return that is determined by a simple rule, like the rate of growth of the labor force plus the rate of growth of the wage. these systems should adjust automatically to demographic change, for example low fert causes slow labor force growth which reduces the rate of return earned on these accounts. And mortality decline raises life expectancy and means that the fund accumulated at the time of retirement, which must be converted to an annuity by the govt, converts to lower benefits per year. I can well imagine a system like that in the US, but it does require some sort of safety net or redistributive program to round out the system.
Marlene Lee: How do the financial implications of population aging in the United States compare with the effects of population aging in other countries? What lessons, if any, can the U.S. learn from institutional arrangements in other countries?
Ron Lee: Well, population aging in the US is much more mild than in other industrial countries because our fertility here is around replacement level, rather than being way below it as in europe and Japan. In my view, we could keep our current Soc Sec system in its present form if we wanted to, by raising the payroll tax by about 4 or 5% which would achieve indefinite sustainability. This is not an option in these other countries because their aging will be so much greater, their public pensions are much more generous than ours, and their ages of retirement are generally younger. Health care is a different matter, and I won’t try to address that although it is extremely important.
Marlene Lee: There is a lot of information on the fiscal impact of programs for the elderly, but what is the impact of population aging on private spending, e.g. does family spending on elderly care mean that they are spending less on education for other family members? Are the effects on private spending similar or different across countries?
Ron Lee: In the US, financial private transfers or actual support for their consumption, to the elderly by family members is not common and averages out to close to zero. But in most of the Third World, family support is the main source of support for the elderly, and indeed pop aging imposes heavy burdens on these families and may compete with their ability to fund their children’s education. In many of these countries, private spending on children’s education is substantial. In the industrial nations, the competition between consumption costs of the elderly and education for the children plays out in the public sector rather than in private spending, and I believe it is a grave concern. I am very worried about what will happen, in the US and in other industrial nations.
Carl Haub: Immigration of people of working age helps keep the worker-retiree ratio more manageable, at least I would think it does. But immigrants age too and would become part of the “problem” themselves at some point. So, it would seem that, with any future restrictive immigration legislation, we would be shooting ourselves in the financial foot. Agree?
Ron Lee: Carl—I agree with your framing of the question: immigration helps to some degree with the old age dependency ratio, but the effect here is much smaller than most people would expect, because of your second point—that the immigrants age too. So in my view, the fiscal effects of immigration are not very important, and I think immigration policy should be set based on other considerations.
Meir Sokoler: Consumption per-capita increases with age, but is it also becoming more concentrated on fewer products and services? Are there any studies on this?
Ron Lee: Meir: A tough question. As you say, consumption per capita rises strongly with age in the US and in many European and other industrial nations, although it is flat with age in most Third World countries (all this in cross-sectional terms). But what is the composition of that rising consumption in the US? A great deal of it is due to a higher proportion of spending on health care, both through private spending and through public spending (Medicare, Medicaid). And after age 80 or so, long term care kicks in dramatically. If we focus on the non-health care and long term care part of private spending, I am not very sure about how the composition of spending changes. A student here at Berkeley in Demography, Emilio Zagheni, did a very interesting research paper on this topic, as part of looking at how pop aging would affect the CO2 emissions, but I don’t recall his findings.
charlie teller: Ron, Please relate your topic to the dual issues of younger-worker immigration to the US, and especially of health care workers from abroad. European countries attract young workers to meet their aging problems, and what have been the financial implications? What can the US learn from them?
Ron Lee: Charlie: Immigration of younger workers has less demographic impact than one might expect on, say, the old age dep ratio in 2050, because they also grow old. To have a big and lasting effect, there must be an accelerating rate of immigration. The fiscal gain is greater if we restrict immigration to those with higher skills or higher education, but I am not suggesting that as a policy because there are many other considerations. The question of health care workers is certainly important, and I am not very well informed on this. But I see it has having two parts: first, there is immigration of less skilled workers who take low paying jobs in nursing homes and hospitals, helping to keep those institutions running at a lower cost than otherwise. Second there are the higher skilled immigrants who have been trained as nurses and doctors. Well, others will know better what the issues are here.
John Gist: Ken Manton has estimated that declines in disability rates, declines in nursing home utilization rates, and improvements in health and longevity will vastly improve the outlook for Social Security and Medicare? How do you assess his findings and conclusions?
Ron Lee: John—Ken Manton has done very important work in this area, and was the first to show this decline in disability rates, which is extremely important for our long term outlook. But the picture now looks more complex, e.g. as discussed in an article in Demography a couple of years ago by multiple authors, perhaps including Manton, Schoeni, Martin, Freeman. They reported that while IADLS, the less severe forms of disability, have been declining steadily, the ADLs, which are more basic and lead to nursing home stays, have not been declining. So I am less clear on what the bottom line is here than I was a few years ago. I want to see more of the international evidence. Also, the role of assistive devices and other new technologies is very important and may make it possible for people to remain active and at home for much longer even if their biological functional status has not improved much compared to earlier people of the same age. This is all very important, not only for its fiscal implications, but also for quality life implications of older people.
Kelvin Pollard: How to you think the incoming Obama Administration might address this issue? Do you have any advice for the new administration?
Ron Lee: Kelvin—important question. 1) I think immigration policy should be determined on the basis of other considerations; I don’t think the fiscal implications of immigration are very important when you combine federal, state and local impacts. 2) Population aging in the US is much more mild than in other industrial nations. Other things equal, pop aging and rising old age dependency ratios will lead to about .2% per year slower growth in consumption per equivalent adult consumer than otherwise. This strikes me as being really quite small. Of course, this effect is concentrated in certain areas, particularly in public pensions, health care, and long term care, and if we focus on those alone the effect is much, much bigger, proportionately. My own view is that we should stick with our current Soc Sec benefit structure, while raising payroll taxes by about 4 to 5%, which would put the system on a sustainable basis for the indefinite future. We could then have some sort of govt managed private program on top of this, perhaps, but without reducing the current PAYGO benefit structure. 3) The really big issues is health care and perhaps long term care. (“perhaps” because of uncertainty about disability trends). I am not going to discuss those.
Dana Hess: As the ratio of elderly to the working-age population in the United States increases over the next few decades, how does this effect employers who provide benefits? Is there any data/evidence that depict an increased cost of covering older workers?
Ron Lee: Dana—a good question, and I am the wrong person to answer it. Nonetheless, I will try to say something here, although this may be wrong. I think that indeed the costs of employer provided health benefits for the elderly is high, and discourages them from hiring or retaining older workers. This exacerbates the problems of population aging, since it makes it harder for older workers to work, perhaps leading to earlier retirement etc. Wish I could say more.
John Haaga: Discussions of the fiscal impacts of population aging in the US highlight Medicare and Social Security, which is natural enough. But long-term care for the disabled elderly is also a big concern—a mixed federal-state-local responsibility. Has there been recent work on either explaining or forecasting expenditure trends?
Ron Lee: John—Good question, and I am not up on what has been done on this lately. Of course the numbers of the oldest old are rising very rapidly, and those are the main clients for long term care. But, as has come up earlier in this discussion, disability rates have also been declining, at least as measured by IADLs, but perhaps not for ADLs. This means (I say optimistically) falling rates of long term care usage at a given old age, but increasing numbers at risk at each age. What this will mean in terms of long term care use per working age person, I don’t know.
Philip Sampson: Are there any studies that show the relationship in an industrial nation between public spending on the elderly [and] spending on education?
Ron Lee: Phil—I only know of one such study, but I bet there are others. The study I have seen is by Gruber and Wise, the same guys who did pathbreaking work on retirement. This study looked at govt spending in OECD countries over a few decades, in relation to proportion elderly. As I recall, they found that for each 1% increase in the proportion elderly in a country, there was .5% increase in public spending on the elderly. This meant that with pop aging, the benefit levels per old person tended to drop slightly (because .5%<1%). But they also found that total govt spending was unaffected, and that means that the 1% increase in proportion elderly, leading to .5% increase in spending on the elderly, was crowding out .5 percent of spending on other things, including education. There have also been a number of studies of the effects of varying proportions of elderly across school districts or larger geographic units in the US and school spending per pupil. I think the results have been inconclusive. But in this case, it is a matter of voting behavior rather than of crowding out in the govt budgets, I think. We found that in California, at the state level older people pay more in taxes than they cost in benefits. Not sure about the local level, but I would expect the same to be true there.
Marlene Lee: How does social spending on the elderly compare with spending on children, not just in terms of expenditures but also with respect to how these expenditures may provide economic stimulus?
Ron Lee: Marlene—in the US, I think around four times as much is spent per elderly person as per child, combining federal, state and local programs. Of course, at the private level, we spend a great deal more on rearing our children than we do on supporting our elderly parents. That is the public/private division of labor in the US and most industrial nations including Japan at this point, but in many countries it is quite different. Providing economic stimulus: Here I am not sure. My first inclination is to say that a dollar spent for an old person has just the same effect as a dollar spent for a child, but I have not thought about this before and could be missing a key point.
Meir Sokoler: In some of your studies you present empircal evidence on the differences between realized and anticipated surviorship rates from different [vintages]. Is there any systematic data base referring to the distribution(s) of forecast errors in this regard?
Ron Lee: Meir: Tim Miller and I published a paper in Demography, maybe it was 2003 or 2004, on assessing the performance of the Lee-Carter method. I think that article has the information you are asking about. Since that time, i think there have been quite a few studies, including a recent one Soc Sec, on this. The Lee-Miller paper looked at imaginary forecasts done using the method starting in around 1920, for all forecast horizons, and then 1921 etc., so there was a huge data base developed. Ron
For additional information:
Recent Trends in U.S. Mortality and Population Aging, an interview with Ron Lee, PRB webcast and podcast.
“Older Workers and Retirement,” in Today’s Research on Aging, , a newsletter produced by PRB and funded by the University of Michigan Demography Center. Ronald Lee and John Haaga, Government Spending in an Older America.
You will find links to Dr. Lee’s publications at www.ceda.berkeley.edu/papers/rlee/.