Bryan Caplan  

The Real Dependency Ratio

PRINT
Miscommunication with Megan... Two Graphs About Dependency...

Most economists will tell you that the real problem with Social Security is demography. (Here's William Poole). The fraction of people over 65 will keep rising, implying a growing dependency ratio.

The problem is that this is only the ratio of retirees to working-age people. But wouldn't a better measure of dependency just be the ratio of workers to people? When you think about it this way, the aging of the population is balanced by (a) the fall in the fraction of children, and (b) the rise in female labor force participation. As one website that I wouldn't normally like explains:

It's the overall dependency ratio—–the number of workers relative to all non-workers, including the aged, the young, the disabled, and those choosing not to work—that determines whether society can "afford" the baby boomers' retirement years. In the 1960s we had 1.05 workers for each dependent, and we were building new schools and the interstate highway system and getting ready to put a man on the moon. No one bemoaned a demographic crisis or looked for ways to cut the resources allocated to children; in fact, the living standards of most families rose rapidly. In 2030, we will have 1.27 workers per dependent. We'll have more workers per dependent in the future than we did in the past. While it is true a larger share of total output will be allocated to the aged, just as a larger share was allocated to children in the 1960s, society will easily produce adequate output to support all workers and dependents, and at a higher standard of living.
I'm not sure I trust their exact projections, but the fundamental point seems sound. One non-working segment of the population - the elderly - is growing as a percent of the population; two others - kids and women out of the labor force - are shrinking. (Though don't forget about the rise of non-working adult men!) Has anyone redone the standard Social Security projections to take account of this?

Update: I found respectable projections for the U.S. worker: population ratio here. Just search for Figure 1. It's still expected to decline in coming decades, but the change is quite a bit milder than the retiree: working-age adult ratio.


Comments and Sharing





COMMENTS (34 to date)
Dan Weber writes:

If a senior citizen is working part-time, does he count that as a dependent, a worker, or both?

I suspect he's treating them as non-dependent workers, but they are really both dependents and workers.

Art Woolf writes:

One problem with lumping together kids and the over 65 crowd is that kids are pretty cheap. They are willing to live in a house with (usually) two older people that are called parents. They do consume a lot of education, and more education than older people.

That's more than offset by what older people consume. They want to live in their own homes, and they consume a lot of health care--a lot more than kids. On balance, I'd say it takes a lot more of working people's resources to support over 65 year olds than it does to support under 21 year olds.

Stan Greer writes:

The problem with Bryan Caplan's theory is that he doesn't think very far ahead.

There's only so much "bang" you can get out of reducing the dependency ratio by reducing the number of dependent children.

Within a couple of generations, you have far fewer available workers than you did before and the only way to reduce the dependency ratio further is to have virtually no kids at all.

Of course, if America completely depopulates itself, that will "solve" the Social Security/Medicare problem, but not in a good way.

For some reason, libertarian economists like Bryan Caplan understand it makes sense for a society to invest in things, but don't seem to see any point in "investing" in future generations of employees/employers.

It's remarkably short term and shallow thinking for someone who fancies himself to be far smarter than the average Joe.

Lord writes:

While the dependency ratio will increase for some time, it will flatten to constant over the next few decades. It is already higher in many other industrial societies than it will ever become here. We are merely transitioning from a growing to a stable population.

Bryan Caplan writes:
Stan Greer writes:

The problem with Bryan Caplan's theory is that he doesn't think very far ahead.

[...]

Within a couple of generations, you have far fewer available workers than you did before and the only way to reduce the dependency ratio further is to have virtually no kids at all.

Of course, if America completely depopulates itself, that will "solve" the Social Security/Medicare problem, but not in a good way.

For some reason, libertarian economists like Bryan Caplan understand it makes sense for a society to invest in things, but don't seem to see any point in "investing" in future generations of employees/employers.

I guess you've missed my long series of natalist posts, Stan. I share your concerns. But I've still realized that the standard dependency ratio numbers exaggerate the problem of dependency.
Bryan Caplan writes:
Art Woolf writes:

One problem with lumping together kids and the over 65 crowd is that kids are pretty cheap. They are willing to live in a house with (usually) two older people that are called parents. They do consume a lot of education, and more education than older people.

That's more than offset by what older people consume.

Look at the paper I cite in the Update, Art. Figure 1 includes the "Weighted support ratio," that adjusts for the smaller consumption needs of children.
8 writes:

Here's the argument in a nutshell: Society gave me money to care for children, but I'm wasting it on other things and now I have to give it back.

Stan Greer writes:

Bryan C., I'm sorry that, forgetting some of your other previous writings, I wrongly lumped you in with certain libertarians of the "people are problems" school.

Of course, there are also a lot of anti-immigration conservatives and birth control-obsessed liberals who fall into that category, too. It's a big tent.

However, although I was wrong to draw too broad a conclusion from what you wrote in this one post, I still think it makes no sense whatsoever to claim that as a society over time you're "lowering the dependency ratio" by having, say, only one child rather than two. Of course, in the U.S. we haven't reached the one-child point yet, and I hope we never will, but your post didn't distinguish between the U.S., Japan and Europe.

Within a generation of starting the one-child-a-couple habit, the dependency ratio will be far higher than before you started "lowering" it -- barring an unprecedented flocking of the elderly into the work force. Although I think more people over 65 probably should work, experience suggests it won't be easy to persuade them to do so in an acceptable, noncoercive way.

Bryan Caplan writes:

Reply to Stan:

Within a generation of starting the one-child-a-couple habit, the dependency ratio will be far higher than before you started "lowering" it -- barring an unprecedented flocking of the elderly into the work force.
You're right that reducing today's dependency ratio by having fewer kids raises it in the future. You can see this effect in Figure 1 of the paper I referenced in the Update. Nevertheless, focusing solely on the ratio of retirees to working-age adults gives an overly pessimistic snapshot of what's going on. Compare Figure 1 to the factoid that the dependency ratio will double during the next few decades. Both are true, if you understand the definitions, but Figure 1 is more meaningful.

spencer writes:

The difficulty that arise under the economic projections by the social security commission do not stem primarily from the changing demographics.
Rather they stem largely from the very low productivity assumptions in the commission projections. If you assume that future productivity growth is similar to past post WW II productivity there is no problem with social security.

You do not have to take my word for this. Check with Tyler Cowen -- he has made the same statement.

Stan Greer writes:

"[R]educing today's dependency by having fewer kids raises it in the future."

With that clarification, I have no disagreement with your point here. I realize your link made this clear, but it seems to me this is a pretty significant qualification of your original point and perhaps belonged in the post itself.

Phil writes:

I wonder what the numbers look like if you (somewhat cynically) take out the number of government employees as part of the productive work force on the grounds that they don't really produce anything but only redistribute others' production.

Randy writes:

Interesting... it means that I'll probably get to collect... not so sure about the kids.

Matt writes:

In not quite serious response to Stans comments:

So, if we have fewer children we can offset the negative trend in dependency ratio with immigration. We could bias our human capital imports toward workers who we think are unlikely to have children and essentially export child rearing to countries where it's less expensive (countries that have a comparative advantage for childrearing). In practice this probably wont work because there is a great potential for people to have more dependent years as seniors than as children so even if we exported all child rearing we'd have to grow the population aggressively to cover the cost.

ed writes:

Bryan also seems to be ignoring home production. Some of what these "non-working" women did is now being outsourced to restaurants, day-care facilities, etc.

Stephen W. Stanton writes:

True enough, but a bit different.

1. You can choose to have kids or a nonworking spouse, but old folks are thrust upon you.

2. Kids are easier to plan for due to finite time horizon

3. There's an equity issue, as the next crop of dependents got a chance to be a burden twice.

Matt writes:

We might say that the dependency ratio will always be within a normal range, for otherwise the system would collapse.

8 writes:

You can choose to have kids or a nonworking spouse, but old folks are thrust upon you.

Exactly. This argument still requires higher taxes. It merely says that people will be able to afford it, which is the argument we always hear from tax hikers.

spencer writes:

phil asks what would happen if you incorporated government employees in the productivity estimate.

That is exactly what the social security commission does. They include government employees -- with the assumed zero productivity -- into the hours worked
data in their productivity calculation. that is why both their productivity history and projections are so much lower then the official BLS data.

Chuck writes:

I think it is worth while to include productivity in the analysis as well - what 1 worker could provide 50 years ago is a lot less than what 1 can provide today.

Victor writes:

Bryan -- The Social Security funding problem is about retirees and workers. Most projections indicate that there will be a need for greater contributions from working age persons to retirees (or a reduction in "scheduled" benefits). I think that your post simply suggests that increased contributions from workers may be feasible because workers will have fewer other dependents and will, therefore, have more discretionary income to tax. Such a position is not contrary to Social Security projections; there is no automatic revenue collection mechanism that will translate the financial "windfall" of fewer dependent children into higher payments to the government to provide for retirees.

dearieme writes:

My immediate reaction was identical to Phil's, wondering what the numbers would look like if you took out the number of government employees as part of the productive work force. Then I realised that that wouldn't do, because it fails to allow for the way they actually impede the productive part of the workforce. Also, we should allow for all the people who appear to be part of the productive workforce but are, in truth, just a sort of government gendarmerie - diversity officers, health and safety Stasi, and so on. Still, your main point is well made.

Dr. T writes:

Bryan Caplan writes:

Nevertheless, focusing solely on the ratio of retirees to working-age adults gives an overly pessimistic snapshot of what's going on.
No, it doesn't. The fact that some workers choose to have fewer young dependents (who usually are known, loved, and planned for) does not make them content with supporting more retirees. Therefore, the ratio of workers to retirees is important to most workers, who are expecting their standards of living to rise. Today's workers, who as a group have fewer children, do not expect or want their extra purchasing power to be sucked up by taxes benefiting an enlarging pool of retirees.

Steve Sailer writes:

Female labor force participation rate peaked in the 1990s and is down since then.

Steve Sailer writes:

Actually, the total fertility rate was higher in 2006 than in many decades. It's just that the kind of people _you_ are friends with aren't having many kids. From 2005 to 2006, the number of babies born to white married women dropped 0.4%, but the number born to unmarried Hispanic women climbed 9.6%!

dcpi writes:

Are the number of children a family chooses to have an independent or dependent variable in relation to the Social Security tax rate? If the two are correlated, do we really want to be replacing children with retirees?

Those who make the argument Bryan cites miss an important point, one that a couple of other posters allude to: people choose to have children, they are literally a good. People do not typically volunteer to pay for non-related retirees (especially at the rate of one-eighth of their income). That implies that Social Security has a significant dead-weight cost.

For those who would argue that the costs of children are socialized also (public school, etc). It should be noted that to some extent the degree to which one bears those burdens can be selected.

Where I live in New Jersey there are very high property taxes that support the local schools. Each Spring "for sale" signs bloom along with the flowers as parents of children heading to college seek to move across the county line to avoid our 2 percent property taxes.

There is no escaping the FICA tax by moving. It is taking what could be a personal decision -- whether to dedicated a large chunk of one's income to support others -- and communalizing the answer. And, few are discussing the unintended consequences of taking that decision out of a family's hands.

Todd writes:

I wholeheartedly agree with dcpi, and feel he did an excellent job of making explicit some of the objections I sensed in earlier posts.

To add my two cents to the distinction between retirees and children, I would note that the latter have no choice but to be dependent, while for a large segment of the former, self-support is a viable option made markedly less attractive at public expense.

By the same token, most of us will some day reach a point as we grow older when we absolutely will need physical assistance to maintain ourselves. It seems to me that this has historically been an important part of the decision to have a family. Though we've probably all doubted it at some point, I imagine that most of us perceive their parents as as much of a good as their children. I'm not radically conservative, but it does strike me that there may be some significant psychological costs that arise from Social Security's interference with this traditional familial relationship.

Johannes writes:

Although I am somewhat unfamiliar with the American Social Security, I think it may be similar to what we have in Canada. In Canada we call it CPP.

I'm less concerned about the change in demographics of the workforce than I am with change (increase) in CPP in terms of the real dollars being distributed on average. Any increase in real dollars creates more problems in a pay-go system as these rents are immediately capitalized out of the system. They only exist for the current consumers of CPP and any reform ex post will be increasingly difficult.

We seem to laud these increases in Canada as great social achievements with total disregard for the limitations it puts on the dynamism of our system.

scary.

Ed Lopez writes:

It seems workers/population picks up a lot more than dependency. For example, would rising aggregate leisure be a good or bad? How about rising labor productivity? Workers/recipients gets us closer to the fiscal strains of OASDHI proper.

Nathan Smith writes:

I think there's a problem with treating the rise in female labor force participation as reducing the dependency ratio.

If you think of housewives as merely dependent, then yes, wives working reduces dependency. But what if you think of them as partners in a household division of labor. That is, a household needs money, and also cleaning and cooking and interior decoration and childrearing and so on. If wives go to work, who does what they used to do? Some combination of (a) husbands and/or wives do it in their after-work time, sacrificing leisure, (b) it gets outsourced to the market-- we eat more restaurant meals and fewer home-cooked ones; we hire maids and nannies, or (c) it doesn't get done at all, which, as long as we assume the housewives were creating value, represents a decline in living standards. (Maybe the loss is in dirtier homes, less-tasty meals, or weaker parent-child ties.)

Presumably housewives wouldn't work if these costs weren't worth it. But the costs of women working still need to be recognized.

Now, if we say, "There are more seniors, but that's okay, because more women are working," the implication is that people will be taking the loss of housewifely services without compensation. That is, two-income couples will be getting higher incomes, and the extra income will be taxed away to support more seniors. That doesn't seem like a reason to be less pessimistic about the rising dependency ratio.

John Thacker writes:

There are two quibbles that I would have with using that dependency ratio:
1) Social Security, I believe, makes up a larger proportion of support of the elderly than governmental programs make up support of children and non-working women (and adult men). This makes it more inefficient economically than supporting relatives out of one's own money. This, and the various higher tax rates required to finance such programs, may affect the projections of the future labor supply by affecting propensity to work.

2) It's certainly accurate that housewives should not be counted in the workers-to-retirees number for the funding of Social Security, but I'm less sure that they should be counted as pure non-workers and pure dependents in the dependency ratio. While the rise in female labor force participation is important, good, and efficient from an economic perspective because it has people at their most productive, counting the economic production of a housewife at zero is inaccurate, particularly if a family hires a housekeeper or nanny to perform work that might have been performed by a parent staying at home. The imputed economic value of staying at home is at least what it would cost to pay such domestic workers.

Surely one can concede that, at the very least, comparing a single parent with a kid to a married couple, one of whom stays at home, with a kid, that the first worker does not necessarily face half the burden of the second, particularly when the child is young. The stay at-home parent can take care of the child and do housework, reducing the resources that the worker has to put forth to support the child.

John Thacker writes:

I see that Nathan Smith got his comment in while I was composing mine. :)

It's beancounter thinking, not economic thinking, to pretend that the economic contributions of housewives are zero.

spencer writes:

The Fed paper is good, but isn't the analysis too static. For example, under these projections labor will become relatively scarce. If so, wouldn't the reaction be to increase the capital per worker ratio so that labor also becomes more productive?

To make long run projections without assuming that there will be secondary reactions to the demographic shifts seems to me to be of questionable value.

We are not sure what the secondary reactions will be, but at least we should try to think about some of them.

Miracle Max writes:

BC's point was first made by Robert Eisner, I forget in which book. "How Real is the Federal Deficit," "The Misunderstood Economy," or maybe "The Great Deficit Scares."

Comments for this entry have been closed
Return to top