April 26, 2005
National Press Club
There was a Dilbert cartoon last week that sort of got my attention and kind of summarized what we want to do today. The pointy-haired boss is having a staff meeting and he makes some statistical assertion and one of his staff challenges him on the assertion by actually asking a question about the numbers, and his response is sort of a nasty response. He says, Now don't get all mathy on me. (Laughter.) There has been a lot of urban mythology about the effect immigration has on Social Security, and we wanted to get mathy on you and actually look at the numbers and see whether this Social Security urban myth that immigration is the solution in fact has any substance to it at all. As you can see from the press release, it doesn't. And we'll go into that in some detail.
The author of the report will speak first. Steven Camarota, to my right, is the director of research at the Center for Immigration Studies, one of the country's top experts on the impact that immigration has on the United States. And then we'll have two respondents.
Nicholas Eberstadt, who is the Henry Wendt Scholar in political economy at the American Enterprise Institute, has become something of an expert on North Korea, which isn't why he's here. He's actually written a good deal on demographic issues as well. He's got a book out. His most recent book was Health and the Incoming Equality Hypothesis, and has written on this issue of aging societies in Foreign Affairs last year, although that was about China and India.
Our other respondent is Lindsay Lowell, who is director of Policy Studies at the Institute for the Study of International Migration at Georgetown, and before that he was the director research at the Pew Hispanic Center, and before that on the staff of the Commission on Immigration Reform. That was Barbara Jordan's immigration commission in the 1990s. So we have an expert panel that should have some interesting responses to the report and to the issue in general of what happens to an aging society and what role immigration plays.
So to set the stage, Steve will start, then our two respondents, and then we'll take some questions from the audience.
STEVE CAMAROTA: Thank you, Mark. I should say at the outset that I do have a bad cold, so if I start coughing uncontrollably you'll know why. It's not that I smoke four packs a day, even though it may sound like that at times.
Now, there's no question that American society is aging. In 2000, the average age of the United States was 36 years, compared to 33 years in 1980. Now, many advocates of high immigration argue that immigration is extremely helpful in dealing with the challenges created by an aging society by making the country significantly more youthful and changing the overall age structure in the United States. Ben Wattenberg of the American Enterprise Institute, a colleague of Nick, to my right, is one of the most prominent thinkers espousing this point of view, but he is by no means alone. Just for example, Washington Post columnist Charles Krauthammer has said that America has been, quote, saved by immigrants, unquote, from the kind of aging taking place in other first-world countries.
Now, the primary worry of such advocates is that in the future there will not be enough people of working age to support the economy or pay for government. This paper that I'm presenting today attempts to answer the question, does immigration actually change the age structure of America, and if so, by how much?
Now, in order to determine the actual impact of immigration I look at four separate areas. First I use Census Bureau data to determine how much immigration has changed America in terms of age. If immigration fundamentally transforms the country in that way, then the record level of immigration in recent years should have some discernable impact. To do this I simply remove the foreign-born population and their children from the 2000 census and then recalculate the average age.
I also calculate from the Census what is called the dependency ratio. The dependency ratio is the share of the population that is either too old or too young to work, relative to the share that is of working age. And the Census Bureau and I do in this paper defines the working age as 15 to 64 years of age. People outside that are considered in the dependent population; those in it are the workers. In effect we try to see what would be the age structure in America with and without immigrants. Now, I mostly concentrate on recent immigrants; I'll elaborate on that later.
Now, in the second area I look at U.S. fertility using what's called the Current Population Survey, collected by the Census Bureau, and with that I do basically the same thing. I look at how many children can be expected to be born per woman in the United States called TFR, or total fertility rate with and without immigrants using the Current Population Survey. Again, if immigration makes us a much younger society, we would expect that immigration has a big impact on fertility, the number of children per woman.
Let me point out that it's estimated that about 90 percent of illegal aliens do respond to Census surveys and the 2000 Census. So in that part of the report I'm looking at basically the impact of illegal immigration and legal immigration together.
Now, in addition I also look at the future. Or I should say I use the work that the Census Bureau has done, in the third part of the study. What they've done . . . in 2000 they issued some population projections in which they looked out over the next century, and based on different immigration assumptions they tried to look at what the overall size of the population would be and what share of the population would be of working age in the course of the next century, assuming, again, different levels of immigration.
And finally, in the fourth and final section I look at some work that the Social Security Administration's done, specifically where it looks at different levels of immigration over the next 75 years to see how immigration impacts the Social Security system. So there are the four areas.
So what do we find? Well, the report itself has a lot of information in it, and I'll try to just hit sort of the highlights here.
First, let's look at average age. The average age of a native-born American in 2000 was 35.4 years, according to the Census. And for immigrants it was 39.1 years. It may be surprising to some but immigrants are actually older on average than natives. This would mean, of course, that if we removed all the immigrants from the 2000 Census, the average age in the United States would actually be 34.4, the age of native-born Americans, which of course is slightly younger than what the average age was, which was 35.8. In other words, the immigrants actually pull up the average age.
Now, let me point out that this does not necessarily mean that the argument that immigration possibly affects the age structure of the United States is entirely wrong. As we will see, the issue is more complex. But it does remind us of an important fact. Immigrants, even those who may arrive very young, grow older with time, just like everyone else. If you will, they have the same problem we all do: they have to celebrate their birthday once a year.
Now, another way to think about the impact of immigration is to examine the effect of only recent immigrants and their U.S.-born children, and thats what I'm going to focus on here.
The Census asks people, when did you come to America?, and you can also since children live with their parents you can simply just take out all the post-1980 immigrants from the 2000 Census and all of their U.S.-born children. If you do that, the overall average age in the United States would be 36.8 years compared to 35.8 years when they are included. In short, the average age in the United States is 36 years when you include the 28 million immigrants or 28 million post-1980 immigrants and their children, and when you take them out it's 37 years.
While average age is only way to think about age structure, the results make clear that immigration in the last 20 years has had a very modest impact on the nation's age structure, even though the last 20 years have witnessed the highest numbers of immigrants in our history.
Let me turn now to the dependency ratio that I discussed earlier. In 2000, 66.2 percent of the nation's total population was of working age. This figure of course includes the immigrants. When I take out recent immigrants and their U.S.-born children, the working age share is 65.9. So let me say that again: it's 66.2 with all the post-1980 immigrants and their children; without them it's 65.9.
This is a very small impact. It must be kept in mind because most of the children who have been born to post-1980 immigrants were under age 15 in 2000 and are therefore part of the dependent population. It should also be pointed out that some immigrants who've arrived in the last 20 years are now over 65. In fact, a significant share are.
So the bottom line from when we look at average age or the dependency ratio and take out the recent immigrants is that although there are 28 million more people in the United States than there otherwise would be and this certainly makes for a much larger overall population and a more densely settled country -- it does not appreciably change the dependency ratio.
Now, as I said at the outset, another way to think about the aging of American society is to look at fertility. Fertility of course is very important when examining aging because children born today to immigrants will be living or to natives, for that matter will be living in America really through the next century. Thus, current births have an enormous impact on the future age structure of America. In fact, they are the primary determinant of it.
Now, in almost every single industrial society Europe, Japan and so forth -- the fertility rate has fallen below that which has traditionally been thought to be needed to replace the existing population. That is 2.1 children per woman. In 2000, for example, the total fertility rate of Europe was only 1.4, but in 2000 it was 2.1 for the United States. It was very low in Europe but the United States is at, or roughly at, replacement level. But the question remains, is immigration the primary reason for this relatively high fertility in the United States? It turns out it is clearly not the reason for it.
If you take the current population survey and pull out all the immigrants and then recalculate the average fertility in the United States, it's 2.0. In other words, with the immigrants it's 2.1; without them it's 2.0. Put a different way, 86 percent of the difference between the United States and Europe would exist even if there were no births to immigrants in the United States. And by the way, I should point out that the reason in terms of births that I look at all immigrants and don't just pull out the recent immigrants is it doesn't really make any difference. Almost all the births in 2000 to immigrants were to post-1980 immigrants. People who arrived prior to 1980 who are immigrants, they're too old in most cases to be having children in 2000. There's a few that came as kids and stuff like that, but generally immigrants have their children in the first 20 years, so I should point that out.
Now, it is true that immigrants have more children on average than natives. Natives have about two children on average; immigrants have 2.7 children on average. But this higher fertility does not have a large impact on the overall fertility rate in the United States. And as we might have guessed, since it doesn't have an impact on the overall fertility rate, it's not likely to have much impact on the age structure in the future. It must be remembered that about eight out of 10 births in the United States are to native-born Americans. Although immigrant fertility is higher, that's just not a large enough share immigrant births are just not a large enough share to change overall fertility.
Now, the evidence clearly shows that at least so far the record level of immigration in recent decades has had little impact on the average age, the dependency ratio, or the nation's total fertility rate. But what about the future? As I indicated, the Census Bureau has done a series of projections in the report actually. Figure 1 shows Census Bureau estimates of the working age share that's 15 to 64 years of age under different immigration scenarios. Let me just touch on what I think is probably the most salient point.
The Census Bureau projections show that if the net level of immigration, legal and illegal, averaged 100 (thousand) to 200,000 people a year between 2000 and 2060 60 years of that the working age share in the United States would be 58.7 percent in 2060. Net immigration, by the way, refers to the number of people coming versus going.
Now, they also say that if net immigration were to average 900,000 to a million a year, the working age share would be 59.5 percent. In other words, with roughly 100 (thousand) to 200,000 immigrants a year, it would be 58.7 percent, but if we had several times that, 900,000 to a million, it would be 59.9 less than one percentage point difference. In fact, it's .8 percentage point difference again, hardly a huge difference.
Now, what is important about these projections is that they show that although there are these very large differences in the components of immigration, if you vary that a lot, you don't produce big differences in the dependency ratio. The reason for this is simple. The U.S. population is already very large, so it's hard for immigrants to change the age structure unless you had many, many millions and we'll touch on that question again later. Second, immigrant fertility, as I pointed out, is higher but it's not that much higher. And finally, of course, immigrants age, just like everyone else, and add to the elderly population.
So far I've just looked at just sort of general demographics. I think one of the biggest questions people have is Social Security. It's a hot topic obviously. And one of the arguments people make is that immigration is extremely helpful to the Social Security system. The argument basically is that adding more workers through immigration will extend the solvency of the program by adding more people who will be paying in. But of course it's important to keep in mind that today's immigrant workers are tomorrow's retirees, and as we have seen both now and in the future, immigration has only a tiny impact on the working age share.
Now, the dependency ratio I've discussed so far looks at the population 15 to 64 relative to everyone else. Now, this makes sense when considering funding for all levels of government because workers provide the taxes spent on both the elderly and children, most notably in the case of children and public education. But education is mostly a state expense while Social Security is funded by the federal government.
But it's not clear that it makes sense to view Social Security in isolation, because after all, if expenses rose dramatically for public education or welfare, then our ability to pay for Social Security would certainly be impaired. Also, it's important to think about that Social Security really is in a separate program in that Congress has chosen to take out all the money that went into the Trust Fund. So far Congress has taken $1.5 trillion out of the Trust Fund or roughly $1.5 trillion to pay for other expenses, including education, defense, welfare, that sort of thing. This money of course has been replaced by IOUs from the federal government in the form of U.S. Treasury bonds.
So, again, thinking about Social Security as the separate program may not make sense, but still, it is common to view it that way, and let's think about it that way. It has a separate funding stream anyway even if that funding goes for everything. So maybe it makes more sense just to examine the ratio of workers, or people of working age, relative to the people of retirement age.
Well, the Census Bureau has done that for us as well. The Census Bureau's low immigration projections, which assume, again, 100 (thousand) to 200,000 immigrants a year, show that in 2060, 27 percent of the 15-and-over population will be over 65, or 65 and over in 2060. Now, their medium-level immigration projections, which assume about a million immigrants a year, show that of the adult population, if we had a million immigrants a year, 26 percent would be over age 65. In other words, 100 (thousand) to 200,000 immigrants a year, the ratio or the share of the 15-plus population that would be over 65 would be 27 percent, but with many more immigrants than that, it would be 26 percent. Again, a very small impact on the ratio of workers to retirees, excluding children even though they do have to be supported.
Now, of course, almost everyone agrees that Social Security has a problem. Whether it's a crisis or a problem is a debate we won't get into here, but there's a problem. According to the 2004 report of the Social Security trustees, tax payments plus that money in the Trust Fund that's just the IOUs, if you take that, there is enough of that to provide funding with a gap of $3.7 trillion over the next 75 years. In other words, expected taxes plus the Trust Fund, plus $3.7 trillion would meet the demands of Social Security. The $3.7 (trillion) is the shortfall.
Now, if one does not treat the Trust Fund as a real asset and there's a debate about this then the actual difference between the program's projected costs would be $3.7 (trillion) plus roughly the $1.5 (trillion) that's out of the Trust Fund. So the actual deficit of Social Security is $5.3 trillion. That's the amount of money we have to come up with either in terms of taxes or borrowing to kind of cover both what we promised in Social Security and the IOUs in the Trust Fund, $5.3 trillion.
Now, how does the Social I just think it's important to get that background out of the way. Now, how does the Social Security Administration deal with immigrants? Well, what they do is they use administrative data to determine the age and sex of newly arrived legal immigrants. They just focus on legal immigrants. They hold the number of illegal aliens constant in their analysis. They then assume that the new legal immigrants added to the U.S. population each year will have average tax payments average incomes and average tax payments from the moment the legal immigrants arrive.
So, in their baseline estimates, the Social Security Administration, or SSA, assumes 800,000 legal immigrants a year. Now, according to SSA, if immigration was cut to zero, the funding shortfall would increase by $346 billion over the next 75 years, a seemingly large number. But it's important to keep in mind that this is equal to only 1 percent of the program's $32.9 trillion projected costs over the next 75 years. In other words, projected costs are $32.9 trillion. The impact of zeroing out legal immigration according to SSA would be to increase the deficit by $346 billion, so it's less than 1 percent. Now, compared to the program's funding deficit of $5.3 trillion, it's just 6.6 percent, even compared to the funding deficit that treats the Trust Fund as a real asset, it's 9.4 percent.
I should point out that a recent study by the National Foundation for American Policy estimated the benefit of 800,000 legal immigrants as $611 billion over 75 years, but that was a mistake. And they have since, by the way, corrected that and posted the right number of $346 (billion) that I've mentioned here.
Now, of course it is extremely unlikely that legal immigration will ever be cut to zero, so it seems to me it makes much more sense to compare changes in immigration that are politically possible, and let's do that. Let's say we cut legal immigration by 330,000 down to 470,000 from 800,000. That's a 41-percent cut. The change in the projected deficit then is only $133 billion again, a seemingly big number, but it's only 0.4 percent of the program's projected costs. It's only 2.5 percent of the funding deficit, and only 3.6 percent of the deficit if you count the Trust Fund as real.
What might this mean for average taxpayers? Well, to make up the difference basically of a 41-percent cut in legal immigration, we would have to raise taxes by 0.12 percentage points, and that comes to about $21 a year for the average taxpayer, or $42, doubling it, if you assume that workers bear the full costs of taxes levied on employers. Put simply, reducing immigration from 800,000 to 470,000 a 41-percent cut has a very tiny impact.
Now, let me point out, though, that there are a lot of questions around SSA's estimate, even those tiny benefits that they find from immigration. And let me give you two reasons why. The Social Security Administration is assuming that legal immigrants will have average wages and average taxes from the moment they arrive. But if their average wages are a lot lower, then obviously their average taxes will be lower because taxes are levied as a percentage of earnings. And in fact there's a large body of research showing that legal immigrants have significantly lower earnings upon arrival. So what the Social Security Administration is doing is saying, well, the immigrants they're going to make just as much as natives from the moment they arrive. And that's how they incorporate the immigrants into their estimates, but a very large body of research from people on all sides of the immigration debate say, look, it takes time for the immigrants a number of years, maybe a decade or more for legal immigrants to eventually try to get to the point where they have the same wages as natives. Social Security completely ignores that and assumes that the immigrants are as wealthy as natives from the moment they arrive.
Now the lower income of immigrants also has other implications for public coffers. There's what's called the Earned Income Tax Credit, which is a cash payment to low-income workers. The credit is specifically designed to refund all or part of Social Security taxes paid by low-wage workers. As the IRS states on its website, the credit was created, quote, to offset the burden of Social Security taxes on low-wage workers. So for example, a family comprised of a husband, wife, and two children, with an income of $25,000 (dollars) a year, would get $2,100 from the credit. Their Social Security tax would only be $1,600 on themselves. And my own research shows that legal immigrants are more than twice as likely as natives to get the Earned Income Tax Credit. But again, Social Security Administration ignores the fact that a very large share of immigrants a much larger share than natives will get their Social Security taxes refunded under the Earned Income Tax Credit, which is what it's supposed to do. The fact that the immigrants get this doesn't represent some kind of moral defect on their part. They simply are much more likely to have lower incomes, they're somewhat more likely to have children, and so they get the credit. But the point is that then that's not money that's flowing in the public coffers. It's money going out of public coffers.
Now overall, it is therefore not clear that immigration actually does create the very small benefits for Social Security that the SSA estimates. But even if one ignores the significant differences between immigrant and native lifetime earnings, tax payments, and EIT receipt, the fact remains that SSA's own projections show that legal immigration has a very small impact. Again, cutting immigration by 41 percent legal immigration according to SSA's projections would increase the size of the deficit by just 2.5 percent. Thus the kinds of changes to legal immigration policy that are politically likely have only a very small effect on Social Security. The debate over immigration should instead focus on areas over which immigration, it seems to me, does have a large impact.
Let me sum up and say that there is no doubt that American society is aging. And one can favor more or less immigration for many reasons, but the available evidence indicates that immigration has not had nor is it likely to have a significant impact on this problem. Americans will simply have to look elsewhere to deal with an aging society. Thank you.
MR. KRIKORIAN: Thank you, Steve. Now Nick.
NICHOLAS EBERSTADT: Mark, thank you very much. Pleased to be here with you, and it's a pleasure to comment on Steve's fine study. It is a meticulous and diligent report, and I commend it to all of you to read. In my remarks I'm going to try to do three things very quickly: first, to tell you that the broad conclusion of Steve's report is incontestably true, secondly, to quibble with a few things about interpretation in the presentation, and finally, to give Steve and Mark some more work for the future by asking some questions which I don't know the answer to but somebody may want to go into.
In broad strokes, I take the conclusions of this report to be, number one, as an arithmetic proposition, the United States needs to rely on immigration less than other OECD societies in Europe or Japan, and two, that under any imaginable world future immigration will not be capable of maintaining the age structure that we have today over the next generation or so in the U.S. As Steve has already pointed out, the United States population is very close to replacement level, anomalously and curiously in comparison to practically every other OECD country, where net replacement levels are typically 30 (percent) or 40 percent below the level that would be needed for long-term population stability. Even without immigration, the United States fertility profile is such that it can almost generate long-term population stability. Now as for replacement migration, what sort of migration flows would it take to the United States to maintain the sort of age structure or the ratio of people 15 to 64 to those over 65? The United Nations Population Division has kindly calculated that for us. And if we wanted to keep to about the 1995 or 2000 age structure in the United States, we'd only need about half a billion immigrants over the next 50 years to do that. The United States population, by those calculations, would be a little bit over a billion in the year 2050. Something tells me that's not going to be what's going to happen. So in broad strokes, incontestably, incontestably correct.
Now for some quibbles. In comparing the median age of the existing immigrant population to the existing native-born population, we run into the risk of measuring apples against oranges or, more specifically, of measuring stocks against flows. If I told you that the median age of people who graduated from college in 1980 today was 35 years, that would not necessarily tell you anything about the age at which people graduate from college in the United States. You see what I mean. And what I think we might better be served by looking at is the actual age structure of people who are coming into the United States when they come in versus the existing age structure of the United States as a recipient society. And the Census Bureau's own studies on native-born versus foreign-born population I think there was a study that came out in 2001 suggested that at that point in time, the median amount of time that immigrants who had come to the United States had spent in the U.S. was about 14 years. So we should bear that in mind when we're comparing the real existing immigrant population against the real existing native-born population, I'd submit.
Similarly, you could have two populations that, by different specific numbers, look identical different demographic measures look identical, but that the structure could still be quite different. Let me give you an example. Again from the Census Bureau report on native- versus foreign-born population, the proportion of immigrant population in the 25- to 54-year age group and this is this thing I've got up there, Steve was a little under 59 percent, whereas for the native-born population, it was about 42 percent. If you took a look at all of the population if you took a look at the immigration ages 18 to 64, I think that would be almost 80 percent in that age group, versus about 60 percent for the native-born. What this means is you'd have a very different ratio of people over 65 in the native-born population from those in the immigrant population a very different ratio of people and potential working ages versus potential recipient ages. The proportion of immigrants is small enough that that doesn't affect the overall mix in the Untied States very much. But they do look somewhat different.
And on the fertility front, actually some of the Census Bureau studies suggest that foreign-born fertility is actually a bit less that Steve's report suggests higher than native-born, about 20 percent higher rather than but these are things that we can, you know, fight out in the nerd trenches.
MR. KRIKORIAN: Yeah it does show a lower one, but the 2000 shows higher.
MR. EBERSTADTL: It's a minor methodological issue of interest to nerds like me, and it can be fought out in the nerd trenches. But the point that I think we might bear in mind is, at this point, about one out of every five children in the United States is born to somebody who was born overseas. And that is a consequential proportion of the births that are occurring in the United States.
The question that Steve raises about the net economic impact, the budgetary, fiscal, and economic impact of immigrants and their children is a huge one and a very, very difficult one to get into. You'll appreciate all of the different complexities that are involved there. I would just note that's an awfully difficult question. One of the great economic historians in the U.S. today is Jeff Williamson up at the Harvard University. And he has done an awful lot of work on the economic impact of migrants in the U.S. and other countries. I asked him if he thought that he could develop a model to indicate the net impact over generations of immigrant flows, and he raised his hands. And if Jeff can't do that, I know I can't.
One final question that I would raise has to do with the health status of immigrants versus native populations. Steve has put his finger on the Social Security question, which is a very meddlesome question, and it's obviously exercising a great deal of our political attention today. There is this other elephant in the room, which is the unfunded health liabilities for the U.S. government. And last time I looked at that, estimates were that those liabilities were actually almost 10 times larger than the Social Security unfunded deficit. Having a bit more information and I do not have this information about the health status of the immigration population might be very helpful to us in a couple of ways: number one, by suggesting whether the immigrant population stands a chance of so-called healthy aging so as to be able to participate in the labor force and later life, and secondly, if the immigrant population is less healthy or more brittle to get a better sense of the sorts of unfunded liabilities that their presence in our society might suggest.
But, congrats. And a very interesting report to read.
MR. KRIKORIAN: Thank you, Nick. Now Lindsay.
B. LINDSAY LOWELL: Good morning. Much of this discussion is very number-oriented. I'm not going to change too much of that. I would start off the same gamut. This really bears reinforcing. Immigration, given what I think is politically feasible, is not a solution to the aging of the American society or Social Security. That's not news to a lot of us, but again, it really bears reinforcing. And why might that be? There's a few different reasons, one of which is it's a tail-wagging-the-dog kind of problem. Immigrants are only 12 percent of the population. You know, you just can't have a big effect on means or medians or percentages when you've got a small part of the population, so that's one thing that's going on here.
The other thing is that American natives already have high total fertility rates, and that's kind of important. When I first started hearing about this issue back in 1989, it looked like American fertility rates were coming down well below two, and they've come back up to about two at this point. So given that, you know, immigrants do have higher fertility rates. How much higher is something we might want to discuss, but the good news is natives already have high fertility rates. And the other thing is that if you look into the future and consider what impact immigrants will have on an aging society, you have to remember that immigrants themselves age. And so, you know, it's a little bit of a dog-chasing-its-tail phenomenon in terms of what impact you might think things will have.
So having said that that is, you know, putting context to this notion of what impact immigrants will have on the age of the society and Social Security in the future I also have a few caveats. They tend to run around numerical critiques having to do with working ages and second generations. But maybe I should also point out that recall also that sometimes small effects are important so that when we look at unemployment rates of 5 to 6 percent, it's only a percentage point difference, but a 4 or 5 percent shift in the age structure can create some of that difference in and of itself.
So the other point is that immigrants do have a younging effect on just the working ages. I think what may be contested is what kind of impact, but the numerical impact is there. Let's also just start for a second and reinforce this notion of replacement migration. This really hit the news just three or four years ago as the U.N. developed a report primarily oriented toward the European audience actually, where things aren't so rosy. And they projected, as you heard, the kinds of populations that immigrants would need to keep the society with a population structure, that is a demographic age structure, that looks just like it does more or less today. And those projections 50 years out, for Europe, basically generated the findings that, you know, 59 to 99 percent, depending on which European country, would end up being the foreign-born and their children, that is, essentially the obliteration of the native population. So replacement at those levels probably is politically unfeasible and probably should be.
Some of the other kind of caveats this particular finding of this particular report generated a whole cottage industry of response. And my colleague at Georgetown, Charlie Keely, made the following kind of notes, which I think are important, one of which is that population projections themselves are unstable. That is, if you look in the past and try to see what happens with projections that were made 20, 30 years ago and more, projections get very hazy 25, 30 years out. There's a lot of dynamics in fertility and mortality rates. And we don't know, for example, what's going to happen in migration. In the case of Europe, there was a lot of migration in the 60s and 70s, and then it stopped, came back on in the 90s. Things are uncertain.
And what will happen with fertility? Fertility is very important. In the European context, fertility is running about 1.2 in Italy and in Spain. 1.2 children per woman that's very low compared to about two in the United States. That could come up. Some countries in Scandinavia in particular are really pursuing pro-natalist policies that try to turn that around. Retirement ages are very important, and when you look at actuarial accounting, consider that Italy has a retirement age around 55. Just bumping that age of retirement up a little bit would solve most of Italy's future problems in terms of Social Security, and you wouldn't need immigration. Women in Italy in the age group 50 to 54 have a labor force participation rate that hovers around 30 percent not quite half of what it is in the United States, for example. So you can see that just changing some basic things like age at retirement, labor force participation rates, can have a much more profound effect when it's 90 percent, 80 percent of the population of natives whose behavior you're changing relative to the smaller proportion who are just immigrants.
But let me add a couple of caveats, one of which is this prime age phenomenon, and this, I think, was also being alluded to a little bit. And this is important because as a society ages, where do your productive workers come from? If you look at the relationship between age and earnings in the United States, you find that people really hit their stride in their late 20s, 30s, and then already by the 40s mid to late 40s earnings start to flatten off and tail off. And what do we think is going on? What economists think is going on is that people become more productive as they gain work experience. There's a prime working age in which they earn the most. And so in any economy, there's a need for these prime age workers who are highly productive, and the immigrants are in fact playing a bit of a role. So for example, one of the reasons that we see immigrants have a low overall age high, excuse me, higher age on average than natives is that only 12 percent of immigrants in the United States in about 2003 were under the age of 20. Twenty eight percent of natives were under the age of 20, so if you take that young population out, average age goes up. And why do immigrants have fewer under age 20? Because they don't migrate as kids. They migrate as adults. In fact, 42 percent of all immigrants are between the ages of 30 and 49, as compared to only 30 percent of natives. This translates, in other words, to the fact that if you look in the prime ages of 30 to 39, 20 percent of all U.S. workers are foreign-born. That's not quite twice the overall share of the labor force, which is about 12 percent. So in the prime age working groups, immigrants are, in fact, a disproportionate presence, and they probably have, in that sense, a younging effect or a productivity effect that's important.
Now Steve and I have discussed this a little bit, and of course you have to question what kind of productivity impact is there. And that's an open question insofar as that immigrants do, in fact, often, if they're low-skilled, tend to have low relative earnings. They tend, by that measure, not to be highly productive. So just the fact that they're a disproportionate share of that population doesn't make them a big boost to the economy, but nonetheless, I would point out that it's important to note that they're there especially as we look into the future. And here's a little bit of on-the-one-hand, on-the- the-other kind of thing. It's important to consider the second generation, the second generation of those people in the United States who were born to an immigrant parent. And if you look at just the average age of all people in the United States, you have an average age of 35.3. If you take out that second generation, the population ages but just to 35.9. So at the present day, the second generation isn't having a huge impact on the average age of the population. And that's certainly the case.
But there will be a change a little bit in the future. Now unfortunately, getting that's a little bit more difficult because there aren't a lot of projections of the second generation available. So I'm going to look just at Latinos to give a flavor of what could be happening in work done by Jeff Passel at the Pew Hispanic Center, where he knows that the second generation Latino population is only 28 percent of all Latinos now, but in the next 20 years they'll become 36 percent. They'll drive about 40 (percent) to 50 percent of all Latino growth in the next 20 years will be in the second generation. There are now about 4.4 million Latino second generationers in the 5-to-19 age group, but by 2020 there will be 9 million almost a doubling. That second generation of Latinos will themselves drive 23 percent of U.S. labor force growth. And much of it again will be concentrated in that 20-to-30 age-year group in the next two decades. So in fact, I think it is reasonable to conclude that, you know, if you look at prime age working group, if you look at the second generation as they come into the labor force, you will see an effect on the prime age productivity working ages with the huge caveat that will these immigrant groups be highly productive, highly educated, really contributing in a highly productive fashion? And at the same time, there's the offsetting effect that their immigrant parents are going to be getting older. And if immigrants are today a disproportionate share of the prime age working group, they will be in the future a disproportionate share of the aging population.
So just to draw some quick conclusions and reinforce my opening statement, immigrants are no solution to resolving the age structure or the aging of the United States in the future, and they're not a primary solution to Social Security. Yes, compared to Europe, they're a plus, but the real solutions lie in the native population and issues like retirement, labor force participation, and whatnot. And that's it.
MR. KRIKORIAN: Thank you, Lindsay. Steve wanted to make a couple of responses
MR. CAMAROTA: Just very briefly. Let me restate. Immigrants do have higher fertility. They're more likely and that information is they do have higher fertility. They are more likely to be of working age than natives, and they're more likely to be in their primary working age than natives. And they have a younger average age when you look at only recent immigrants. But it doesn't matter. If you take the 2000 Census, and you (audio break, tape change) use the Current Population Survey or the Census, it's just not enough to make any difference. And that's the meaningful thing if you're going to use immigration to change the country's age structure, it has to have an impact on the age structure.
And one more point, I agree with Lindsay projections are very iffy things. And he's right if you find a four- or five-percentage-point shift based on the level of immigration or in the percentage of workers or something like that but we don't find that in the projections. Even with their very high immigration projections, don't find a shift of that size. And I guess my final point then is if we come back and don't look at the projections because of their difficulty in doing them, if we look at the actual data collected in 2000, or even since then, they don't rely on speculation, the impact of immigration is very modest. And I think what that tells us is it's going to be very modest in the future. Now that doesn't tell us what level of immigration we can have. I think in the end, the point is, you could favor a lot less immigration secure in the knowledge that it wouldn't have any real impact on Social Security and the aging of American society or you could favor more if you like.
MR. KRIKORIAN: Okay, thank you. Any questions from anybody for the audience? (Pause.) Yes.
Q: Steven, you addressed immigration and Social Security, but could you briefly discuss the role illegal immigration plays in Social Security? There's been conflicting things said about what role maybe . . . what role do illegal immigrants pay into Social Security?
MR. CAMAROTA: Some people point out that illegal aliens do pay into Social Security. About half are estimated, by our work, to be paying Social Security tax. And we've estimated that the amount of money that illegal aliens pay into Social Security is about $6 billion dollars a year. And so people say, well see, Social Security can be saved by immigration or illegal immigration. But that's not correct. Because what we also found was that in programs other than Social Security, illegal aliens imposed a burden of $16 billion for a net negative for the federal budget of $10 billion.
So yes, they should illegal aliens do create a net fiscal benefit for Social Security, paying about $6 billion (dollars) a year in Social Security taxes. But they create burdens in non-Social Security programs of $16 billion (dollars) for a net negative for the federal government of $10 billion (dollars). Thus, any suggestion that illegal immigration is fixing or saving the social security system is incorrect because there's the non-Social Security part of the federal budget, which you have to consider.
Finally, it should also be pointed out that although we estimate that illegal aliens pay about $6 billion (dollars) in Social Security tax, that comes to only about 1 percent of all the Social Security tax collected in the United States a very tiny fraction of the total. In other words, we could do without all the illegal aliens and all the Social Security taxes they pay, and it would have basically no impact on the Social Security system.
MR. KRIKORIAN: Next, yes, in the back?
Q: (Off-mike) -- the net negative $16 million (dollars) that you're talking about, what programs are you including?
MR. CAMAROTA: Yeah, that was based on a study that we did over the summer. The largest
MR. KRIKORIAN: Which is on our website, if you want to look at it.
MR. CAMAROTA: On our website. And I think I have a copy I could give to you. The largest costs associated with illegal immigration are health care providing services to the uninsured totals over $2 billion dollars a year. Also, providing Medicaid to pregnant illegal aliens and also the U.S.-born children of illegal aliens a very large percentage are enrolled in the Medicaid program. Other very large costs include the federal prison system about 20 percent of illegal aliens are in about 20 percent of the federal prison system is illegal aliens.
The other big programs in terms of welfare that illegals use is not the cash-assistance programs. Almost no illegals use those. The big programs are the food-assistance programs, namely the WIC program Women, Infants, and Children food stamps, and free school lunches. There are some other costs also that we look at. But we also estimate non-Social Security tax payments. The figure I gave of net $16 billion includes billions of dollars that they pay in other taxes like, you know, in gasoline and in excise taxes on cigarettes, as well as income tax. Illegals though, because their income is so low, don't pay a lot in income tax, but they do tend to pay a lot in Social Security tax because it's a flat tax. Everybody has to pay it, no deductions.
MR. KRIKORIAN: Yes, sir.
Q: Again, a question for Mr. Camarota in your research and your work, you see a significant difference in structure between (inaudible). For example, from Russia and Eastern Europe (inaudible).
MR. CAMAROTA: There is. I mean, do you want to answer that one, Lindsay?
MR. LOWELL: Well, certainly, there is. The biggest differences tend to be the Latino population at large. Again, the estimates there are that, for example, the fertility rate is around 3.5. The fertility rate for Mexicans in the United States Mexican-born individual women in the United States is actually higher than what we believe the fertility rate is in Mexico, at present. So that tends to be the higher fertility rates. Europeans tend to have fertility rates that are more like the native-born population, and Asians, I would say, are somewhere in between.
MR. CAMAROTA: Yeah, Asians are about two as well, for most Asian groups.
MR. LOWELL: The black immigrant population primarily of African and Caribbean origin in the United States, is about 2.5. Asian immigrants have a total fertility rate of about 2.6.
MR. CAMAROTA: Well actually, I think we disagree a little bit on the Asian numbers, because I was trying to calculate oh sorry, it's not a big deal but overall, I estimate immigrant fertility at about 2.7. Hispanic is higher, other groups tend to be lower Europeans and Asians are lower. So you get an average fertility about 2.7.
MR. EBERSTADT: The Census Bureau suggest that the fertility level for so-called White non-Hispanic ask me about that immigrants, which would be mainly European Eastern European immigrants, is about 2 about 4 or 5 percent higher than for the so-called White non-Hispanic native-born population. And as Lindsay and Steve have said, all of the real action is with Latino immigrant population where levels are higher curiously higher than from some of the sending countries.
MR. CAMAROTA: Yes, in sending countries, it's an interesting question.
Q: So European immigrants actually are the lowest in terms of (inaudible)?
MR. EBERSTADT: Among I don't actually know as a matter of fact about Asian immigrants, because I haven't seen that broken out
R. CAMAROTA: Canadian immigrants also have low fertility in the U.S.
MR. LOWELL: Again, we would differ the TFR I've seen for Asians is about 2.6. But there's also huge differences in age structure. So European immigrants, by and large, tend to be much older than Latinos, who tend to be actually a little bit younger than Asians. But again, that really varies strongly by there's a lot of different kind of Asian origin groups.
MR. KRIKORIAN: Well, and I was actually going to make that point. Even within these larger groups, I mean the
MR. LOWELL: There'd be differences there.
MR. KRIKORIAN: The Vietnamese or Cambodian immigrants may well have very high fertility where as Koreans may have lower. So even within the larger groups, there's going to be significant differences. Any other questions? Yes, sir.
Q: I have two questions about total fertility rates. One is, is that for immigrants once they enter this country or is that for their entire lives?
MR. CAMAROTA: It's based on basically children that they've had in the previous year. The way you calculate total fertility is you look at each age group and see how many women in that age group have had a child in the last year. And so basically, what we're measuring is children that they've had in the last year and since the vast majority have been in the U.S. for the last year, their births are in the U.S.
MR. LOWELL: That's actually a very interesting question. It gets to a technical issue about what's a stable population and present equilibrium and all those kinds of odd terms. But for example, there is work that suggests that over their lifetime, non-Hispanic white women in the United States will have higher fertility rates than 2.0 than what we're seeing currently and that for some Latino groups, it may actually be a bit lower than what we're seeing in terms of current numbers.
Q: And then the second question I was going to ask on fertility rates. I know you said hard-to-find information on second and third generation Americans, but if it seemed to me, if you compared only immigrants foreign-born to native-born, and you talked about taking out the foreign-born, isn't there a chance that that would decrease the total fertility rate for the native-born? Because the second generation of Americans isn't necessarily going to all of a sudden drop to the same fertility rates as the eighth generation of Americans, right? Or is there a sudden drop?
MR. CAMAROTA: Well, here we're just looking at I mean, you can do that with the data. Now if you take into account second generation, obviously, one of the problems is that the second generation, the people who are adults today whose parents were immigrants are not the children of today's immigrants. They are the children of yesterday's immigrants. They have a different they're from different countries, different ages so it's kind of hard to say, but it looks like the fertility of second generation Americans is roughly the same as it is for natives. Because they make it's just not a big difference, though maybe in the future that could change because if Hispanics say Hispanics make up a small fraction of native-born people in their childbearing years right now. But maybe if they have persistently high fertility, then they will change the native fertility when they the U.S.-born children of Hispanics reach adulthood.
However, the overall trend in recent years is for convergence. Immigrant fertility has declined in recent years, second generation Hispanic fertility has declined in recent years, Mexican fertility has declined, so my guess is the difference will not continue to narrow over time. And that's why I think that in some ways, the Census Bureau projections are a little off in terms of immigration. Even that small impact is probably not going to be there because I would be willing to bet that immigrant fertility will fall even more.
MR. KRIKORIAN: I actually had a question that is kind of tangential, but sort of related for anybody. Anybody have any ideas on why native-born fertility is higher than in Europe or Japan? In other words, what is anomalous about the United States? I have ideas myself, I was just wondering if anybody had thoughts on it.
MR. CAMAROTA: Well it used to I mean one reason would be that African-Americans tended and they were overwhelmingly obviously native-born tended to have higher fertility. But by now, their fertility isn't that different than whites. It does seem to be associated with to some extent with the religiosity. Evangelicals, conservative Catholics, and Jews tend to have higher fertility and that helps pull up the average and Mormons too obviously. Well, just the highest fertility in the United States is found in Utah. That's why I point that out as a statement. So anyway
MR. KRIKORIAN: Any ideas?
MR. LOWELL: I don't think anybody has a good idea. This has been again a small cottage industry in the demographic community. There was an important paper written in the late 70s that basically said there was an end to Catholic fertility in fact, it wasn't higher. And it looked like there was a lot of downtrend in fertility in the 80s. It has turned around and we're not really sure. That's my take on it. There's a lot hypotheses about religiosity in the United States. There's something about a greater flexibility in the family structure in the United States. But nobody really knows for sure.
MR. KRIKORIAN: Yes?
Q: You've taken a tangential question (inaudible). If there is a large-scale legalization program that Senator Kennedy and (inaudible) are about to sponsor has anyone thought about what the long-term say in twenty, thirty years the demographic implications of that would be, given some integration into immigration how that would flesh out over the next decades?
MR. CAMAROTA: Well, for one thing, the projections assume the illegal aliens are here. Census Bureau just assumes a level of illegal immigration. It doesn't assume that they go away. So they have incorporated illegal immigration into their projections, so in that sense. But obviously, you would probably spur a lot of legal immigration if you legalized the illegal aliens here, because then they would be in a position to sponsor their family members. But it's not clear exactly that there would be a huge impact on the age structure again either way from legalizing illegal aliens.
I can tell you this I've taken out the illegal aliens from the 2000 Census, and what that shows is that without the illegal aliens, the working-age share in the United States is essentially unchanged. With illegal aliens, 66 percent of the population is of working age. Without illegal aliens, 66 percent is of working age. The reason for that is that although the vast majority of illegal aliens are of working age though I should point out, about 2 million are not they're too old or too young is that they only make up about 3.5 percent of the total population, so it doesn't really change the age structure of the United States. Illegal immigration doesn't have a large impact on the age structure, even though the illegals themselves are young, they're not sufficiently numerous and they're not sufficiently different from natives to have much of an impact.
MR. KRIKORIAN: And just to you had asked about sort of what would the kind of cascading effects of legalization be, as far as, being able to then sponsor other people we haven't done any work on it, but I think Numbers USA has done some estimates of the future effects on legal immigration flows of current legalizations. You'd have to ask them, but I think that they've done something on that.
Anyone else? No, okay. Well let's wrap it up then. You can accost the speakers afterwards unless they want to run out of the room screaming to avoid that. And let me thank you all again for coming. Again, the report is on the site, it's cis.org along with all of our other studies. And we'll see you next time. Thanks.