The Age-Old Problem of Old Age

Dependency is among other things a concept/statistic in demographic economics, defined (roughly) as those who are either too old or too young to work. It aggregates two groups: Children, who, because of the evolutionary biology of human development go through a long period—by convention, age 0–14—during which they are unable to take care of themselves. And old people—by convention, age 65 and up—who get to a point where they can no longer take care of themselves. In each case, where taking care of yourself means, more or less, being able to work for a living.

Obviously, real life is messier than this neat 0–14/15–64/65+ trifurcation—in the West at least parents generally pay for children for a longer period and a meaningful number of individuals work past age 65.

Moreover, while the period of childhood dependency is well-bounded, that is, the age at which a child becomes un-dependent has some stability, the period of old-age dependency keeps getting longer, as we discover new ways of preserving life.

And the challenges presented by childhood dependency are different from those of old-age dependency. The young can’t provide ahead of time for their dependency. But the old can and in many cases do.

There’s also a reciprocity between the young and the old here, arcing through time. We start out with parents caring for their children. And as these children become adults and their parents become old, we wind up (in many cases, and in different ways), with children caring for their parents.

Thus, in my view, old-age dependency and the issues it presents are joined at the heart with the dependency of children. And, even in an age where the overwhelming majority of old people live in-dependently, this issue of dependency remains at the center of retirement policy.

Let’s abandon abstraction for a moment and consider dependency as (we might imagine) it was once lived. Let’s think about a small and simple community—a little tribe with (virtually) no economy, no markets, no money system, with little division of labor. For such a group, the issue of dependency was not an abstraction or a statistic, it was a brute and concrete problem: a challenge built into human biology.

They had children who could not take care of themselves, who could not do the basic tasks it takes to produce food. And someone—some productive adults—had to devote some part of their productivity to taking care of—providing for—these children.

Similarly, these same adults, as they aged generally entered a stage of dependency before death. A period—possibly a very long period—during which they were alive but could not feed or shelter themselves. When, again, they had to be taken care of by productive members of the group. In this little imaginary tribe, perhaps by their own children.

The dependency challenge that faces this little pre-economic group is that its productive members, somehow, had to produce enough—e.g., hunt for and gather enough food—not just for themselves but also for their children, and their old people.

Again and again as we consider the issues that old age dependency presents, there are practical, economic and (even) ethical dimensions to this challenge and to our response to it.

In the practical dimension, this challenge means that those who are productive must divert some of their “wealth” to individuals who are not productive. And must forego some other—perhaps even more productive—activity to do so.

Indeed—and this is a feature of retirement policy (and politics) that will be with us all the way through time—if the band cannot take care of its young, it will not reproduce, and there will be no one to take care of its current productive members, when they get old. And, if the current generation doesn’t take care of its own dependent old people, then that (kind of) sets a bad example for their own children.

Thus, while it is theoretically possible for the current generation to simply forego its obligations to its dependents, doing so will doom the group. For this reason, even at the most primitive level, humans have developed ways of addressing this issue of dependency.

While I’ve styled these as practical challenges, they obviously have an ethical dimension. Human compassion dictates that—if we can—we take care of the helpless. But (no doubt) the evolutionary psychologists would argue that “human compassion” here is just an “ethical” way of describing the solution to the practical problem just discussed: if you don’t take care of your dependents, you will die off.

There is, however, another ethical issue—what I believe is properly called an issue of distributive ethics (or distributive justice). Who should take care of the young and old? For a very long time, the answer, at least in the first instance, was that parents should take care of (young) children and (adult) children should take care of (older, dependent) parents. In addition, there has been (perhaps) a sort of residual obligation on the part of more distant relatives, or the group as a whole: for instance, in the case of an orphan or a childless older member.

This link—parents take care of children and then children take care of parents—establishes a sort of rough reciprocity of dependency/obligation that has (implicitly) informed how we have thought about “retirement,” basically forever.

And it is what—in the abstract—puts considerations of demographics at the center of retirement policy.

In this regard—and this will come to matter a lot as time goes on—distributive ethics becomes an issue (is foregrounded, is a wheel that begins the squeak) when either (1) some group (or individual) finds some way systematically to shift their dependency obligations (to take care of their children or parents) on to others or (2) the dependency obligation becomes somehow disproportionate.

The latter point is a little obscure and best illustrated in concrete terms. Consider our little group, where parents take care of children and then children take care of parents. The burden of taking care of a parent where there are, say, four children to share it, is substantially less than the burden of an only child, who must bear it entirely himself.

It must be emphasized that none of this is black and white. Children have some value—they can help with some tasks—and so do old people—they have experience and (hopefully) wisdom that can be invaluable. But, at the margin (as the economists say), at some point the aging process will make them a net drain on the group’s resources.

When this all works—when parents take care of children and children take care of parents—the “system” operates, as it were, in the background, unnoticed. It is just what people in this small group do.

When it breaks down—when there is an orphan child or (more relevant to our inquiry) a childless parent—then at the level of the group, something must be done.

When Ahab dies hunting mastodons, and our little group, some evening, gathers round the fire to discuss what to do about Ahab’s mother, they are talking about retirement policy.

And the way this problem was solved—the only way it could be solved—was to transfer wealth from those who could work to those who (because of old age) couldn’t.

The Wealth Transfer Paradigm

This system—of parents taking care of children and then children taking care of parents—endured for millennia. Its durability depended on two long-enduring realities. First, and most importantly, for most of human history, our demographics have been generally stable. Figure 27.1 shows the world population from 10,000 BC to the year 2100. (We’re going to be coming back to this data a lot in what follows.)

As we can see, and as we just said, for the period from “the dawn of time” to around 1850, world population was relatively stable.

Second, during this time, the period of old-age dependency was relatively short. In the United States, in 1900, life expectancy at birth was 47 years.58 If (as is conventional) we assume old-age dependency begins at age 65, then even in 1900, in first world countries, few individuals actually reached old-age dependency, and those who did lived a relatively short time.

The Socialization of the Paradigm

Then, beginning in the mid-19th century, two things happened. The first thing that happened was that, beginning in the West, nations began to “socialize” the parent-child wealth transfer paradigm. This actually occurred to some extent at both ends of the dependency continuum. That is, things like state financed education “socialized” the cost of child-dependency, in effect making “everyone” pay for the education of children.

At the other end of the continuum, social insurance programs—arriving somewhat late in the United States in the form of Social Security—“socialized” the cost of old-age dependency.

For purposes of what follows, we need to be absolutely clear about what these old-age social insurance programs were. These were not savings programs. Instead, they were nearly universally financed on a “pay as you go” (PAYGO) basis: the current working generations were taxed (in the US, the Federal Insurance Contributions Act (FICA) tax), and the proceeds of that tax were paid old-age dependents. And any surplus produced by this transaction (any excess of current taxes over current benefits) was, in an era of debt financed fiscal policy, simply spent on other things.

Thus, old-age insurance was just a social version of the age-old child–parent dependency paradigm.

Riding the Population Wave

The second thing that happened was an extraordinary burst of population growth. Figure 27.1 shows a rapid increase in world population beginning around 1850. This was a new thing in the world. There were at least a couple of reasons for it.59

First, the industrial revolution made it possible to support a much greater number of people—the human race, in effect, broke out of the Malthusian economics of the deer herd.

Second, a revolution in public health drove a dramatic drop in infant and child mortality, making it possible for many more individuals to survive to adulthood. This set up an interesting dynamic that is still to some extent operating in the developing world today. Parents, habituated by millennia of experience to having, e.g., 4 children just to have a chance at having 2 survive to adulthood, found after this public health revolution that all 4 children survived. It turns out it takes a while for families to adjust to this fundamental change in the infant survival rate. And during this lag, population booms.

If we look at this change in the rate of population growth in Figure 27.2 we see a boom lasting from about 1850 to the late 1900s.

This fact of life made the cost of PAYGO social insurance systems (and debt financed spending generally) largely painless. The future always provided more individuals over whom to spread the burden of current spending.

Most importantly for our purposes: these social insurance programs did not involve any saving. I want to re-emphasize what is, for purposes of our understanding of the retirement policy issues we are discussing in this book, the critical point here: from the dawn of time right through the first half of the 20th century, retirement policy did not involve any saving. It only involved a transfer between generations.

PAYGO Vulnerabilities

This sort of system has always—again, since the dawn of time—had one pretty obvious (and potentially tragic) vulnerability: if the succeeding generation is significantly smaller than the preceding one, it will break down. When the preceding generation ages and can no longer work, there simply will not be enough productive individuals in the succeeding generation to take care of them.

For the period 1850–1980, of course, we saw the opposite of this problem: each succeeding generation was much bigger than the preceding one, making the operation of PAYGO-financed old age social insurance programs deceptively smooth.

The socialization of this system in the 19th and 20th centuries also exacerbated another vulnerability of this system.

As we said at the beginning of this section, under a system in which the succeeding generation takes care of the preceding one (writ small, in which children take care of their parents), if everyone has the same number of children (or at least pays the child-raising costs of the same number of children), then the PAYGO burden is evenly distributed.

In these socialized systems, however, the burden of not having and raising children and therefore not having anyone to support you in your old age doesn’t fall on the individual. That is, under these socialized systems, if an individual has no children, instead of risking old-age poverty, someone else’s children will wind up financing this individual’s old age.

But, for the period 1850–1980, the burgeoning population obscured both of these vulnerabilities.

And Then Everything Changed

As the chart above shows, however, the boom in the world population growth rate reversed in the late 20th century. One result of that reversal is an intense stress on states heavily dependent on PAYGO retirement systems—from Greece and France to the state of New Jersey.60

This change had a couple of consequences that we have been discussing, one way or another, throughout this book.

First: this change in world population growth was, necessarily, associated with a long-term decrease in interest rates. The driver here is associated with productivity, so let’s focus on changes in the growth of the working age population. In Figure 27.3, we map US interest rates from 1955–2015 against the rate of growth in working age (15–64) world population for the same period.

The fit here is uncanny. And, one could track the same pattern in world population as a whole (as opposed to just working age population). The relationship of interest rates to the rate of population growth would just lag by 10–20 years—more or less the number of years it takes a newborn to become a productive adult.

Why this correlation? Classically, as populations grow at a slower rate (or, indeed, in some nations, begin to decline), gross output—classically, gross domestic product—necessarily declines. And, as the population ages—especially in the rich, developed world, which accounts for an overwhelming piece of the demand side of the interest rate equation—the demand for and cost of certainty in returns (specifically, high quality interest bearing securities) goes up.

Thus the problem of interest rates, which, as we discussed in Part I, doomed the DB system, and, as we discussed in Part II, made the issue of 401(k) annuitization such a “bad deal,” wasn’t some accident of the markets. Or even a consequence (in anything but the very short run) of Federal Reserve policy. It was a function of the basic demographics of the human race that we are currently living out.

And just to be clear—the numbers we are talking about are world population numbers. As Figure 27.2 shows, things are only going to get worse. As the under-developed world (at this point, primarily sub-Saharan Africa) figures out that improvements in health will mean that most of their children will survive to adulthood, the third world baby boom that the developed world has depended on to compensate for declining birth rates will come to an end.

Money versus Time

All of these factors explain, at a deep level, the transformation in US retirement policy that we described in Parts I and II. The DB system, emerging in the 1950s and followed in the 1980s by the 401(k) system represented solutions to this demographic challenge.

For millennia the problem of old-age dependency was financed via a transfer of wealth from productive members of the family, group, and nation. But beginning with the implementation of robust funded retirement programs, that is, retirement savings programs, workers began financing their own retirement.

This required the “construction” of a number of institutions that, in Parts I and II, we have described (as it were) from the ground up:

Turning Savings into Investment

Indeed, and particularly with regard to the development of the mutual fund industry, it’s fair to ask which way the causal arrow has pointed: without an efficient way of investing (what is now) $27 trillion in retirement savings assets, it’s unlikely there would be $27 trillion in retirement savings assets.

Indeed, to a large extent, merely saving for retirement does not solve the old age dependency challenge in a context of declining population growth. No matter how much the current generation saves, there will still be the same number of individuals in the future to provide them services. In a context of declining birth rates, greater savings will only translate into higher prices for those future services, as retirees compete to buy them. In such a context, the only way to increase the amount of future services available to the current generation when it retires is by increasing (future) productivity. And that necessarily involves not merely saving, but investment, and specifically investment in productivity-increasing innovation.

And to accomplish that you need an efficient financial sector—capital markets that with minimal transaction costs can turn retirement savings into productive investment.