Social Security and Immigration: A New Way to Model the Impact

A new federal budget modeling tool developed through the Penn Wharton Public Policy Initiative is being unveiled for public use today. Already it is producing new thinking around two areas very high on the public agenda: Social Security’s funding shortfall, and the economic costs and benefits surrounding immigration. In this interview with Knowledge@Wharton, Kent Smetters, Wharton professor of business economics and public policy, explains how none of the common fixes suggested for the Social Security shortfall will close the gap by themselves, according to the new simulator, and that the impact on the economy from providing a path to citizenship for undocumented workers is basically an economic wash for the U.S. The Penn Wharton Budget Model is a nonpartisan, online, interactive budget-modeling tool available at no cost to users. A demonstration of how the model works appears at the end of the video, or here for Social Security and here for immigration.

An edited transcript of the conversation appears below.

Knowledge@Wharton: Kent, there’s a lot of talk that Social Security just won’t be there for younger people. Will millennials receive the same Social Security benefits as their parents?

Kent Smetters: Under current law, the answer is no. In fact, we’re projecting that the trust fund will be exhausted by 2031 in just 14 years. So what the law actually says is that benefits across the board have to be cut in order to match the current payroll tax revenue that they’re getting at the time. That’s about a 30% reduction.

Knowledge@Wharton: If nothing changes then, Social Security benefits payments, monthly payments would have to be cut by 30% for everyone?

Smetters: That’s the key. It’s not just for new retirees, which could be a little less painful. But your 90-year-old grandmother, who’s maybe just surviving on Social Security, the benefits would be cut for her as well.

“If you increase [the Social Security maximum tax] from $118,500 all the way to $400,000 … it basically moves the trust fund exhaustion date from 2031 to 2036.”

Knowledge@Wharton: Of course, everyone’s scrambling to figure out what are some things that can be done about this. One of the common ideas is to raise the retirement age for Social Security. How much does that help?

Smetters: It doesn’t help much for the short term. In particular, it barely moves the trust fund exhaustion date from 2031. The reason why is that, in the current law, Social Security retirement age is creeping up to age 67. Even if you push it out to age 70, this idea that’s been batted around for a couple of decades in Washington, we just kind of waited too long to do it. And so it barely moves the trust fund exhaustion date. Over the long-term, it actually has a much bigger impact. But for the immediate term, on the Social Security trust fund, it barely moves it — simply because we waited too long to do it.

Knowledge@Wharton: What’s the difference between the immediate and the long-term? In other words, at some point we’d be better off, if not fully whole. But how long would the lag time be?

Smetters: The trust fund exhausts in 2031. Because we would phase in toward age 70 along the current path, which is about two months per year, it takes too long. So the trust fund barely moves. Over what’s called the 75-year window, which is how Social Security actuaries typically look at this, it helps improve the finances by almost 1/3. So it actually has a big impact over time. It’s just that it takes a long time for it to phase in.

Knowledge@Wharton: And the other solution often put forth is, “Well, how about if we increase taxes on people with high incomes?” Because there’s a cap right now on how much people pay.

Smetters: That’s right. In particular, you pay taxes up to the first $118,500. That’s indexed with wages, so it kind of grows over time. We call that the tax max. It’s the maximum income that your taxes are levied over. And so if you increase that from $118,500 all the way to $400,000, for example, it basically moves the trust fund exhaustion date from 2031 to 2036. Going above $400,000 has almost no impact, because there are just not enough people there. In fact, going from $250,000 to $400,000 has very little impact. So again, it can be part of a bigger solution, but by itself, it’s not going to solve the problem.

Knowledge@Wharton: What about reducing benefits for high-income people, who presumably don’t need the money as much as that 90-year-old grandmother does?

Smetters: That’s another idea that’s been discussed in Washington. A lot of people don’t understand that the benefit formula is actually extremely progressive. For poor people, we replace 90% of their average income before retirement. And that decreases quite a bit for middle-income and then higher-income households. So the idea has been, “Let’s keep poor people at a whole 90% replacement rate, but lower it for middle-class and upper-income households.”

Even a very aggressive, across-the-board change — for the trust fund by itself — it’s basically too late. It barely moves the trust fund. As far as the 75-year solvency, it again has a much bigger impact. In fact, if you combine that with increase in the retirement age, we actually solve the 75-year problem with those two combinations. However, we still would have to come up with a mechanism for the trust fund to somehow borrow against its future, essentially be able to not just hold assets, but be in debt for a while.

Knowledge@Wharton: You are saying there would be this interim when there would be a shortfall?

Smetters: There would be a shortfall. Depending on the exact policy combinations, it could be between 10 and 25 years.

“It turns out that the impact on the economy from legalization [of undocumented workers] is basically a wash.”

Knowledge@Wharton: So it might involve some [political] bartering….

Smetters: That’s right.

Knowledge@Wharton: The other item is the cost of living allowance, or COLA, that is provided each year — although there wasn’t one last year — based on inflation. And if that were adjusted, how much would that save? Of course, that’s cutting benefits if you cut that.

Smetters: It is. Under current law, what happens is that the Social Security benefits are adjusted every year for use in what’s called the CPIW [Consumer Price Index for Urban Wage Earners and Clerical Workers], which was just what they happened to choose at the time. And one of the concerns that economists have had is, that index doesn’t account for what’s called “substitution,” where you might have a product price that goes up on you, and you substitute toward lower prices. We’re seeing this, for example, with the Affordable Care Act. As a lot of plans get more expensive, people substitute toward cheaper plans.

And so a true measure of inflation would allow for that substitution. That’s called a chain-weighted index. That has been something that has been discussed a lot, ever since the Boskin Commission, and other groups. It actually turns out it doesn’t have that big of an impact. Even though in the long-term, it has a positive impact on the trust fund. Not that big. And another concern that people sometimes raise is that it’s not even clear that even the chain-weighted is actually the right measure for older people. In particular, older people have a lot of out-of-pocket medical expenses that are not representative of the average person in the economy that the chain-weighted index cares about, or even the CPIW, the current law. That will be something that there is going to be a lot of discussion about. In terms of the numbers, it doesn’t really have a big impact.

Knowledge@Wharton: If Americans, together, paid more taxes, would that save Social Security as we know it?

Smetters: Sure. If you increase taxes or cut benefits, you’re going to solve the problem. So if we think about the tax side, right now we currently levy a payroll tax rate of about 12.4% for Social Security, and also the disability program. And then an additional tax to cover Medicare. So if we increase that 12.4% to say, 14.4% — that’s two percentage points….

Knowledge@Wharton: It’s shared by the employer and the employee?

Smetters: It’s shared by both sides. Now, truth be told, most economists believe that ultimately, the employees bear the brunt of even the employers’ side — that the increase gets shifted through competitive labor markets. If we did that, if we went from 12.4% to 14.4%, we would increase the life of the trust fund from about 2031 under our projections, to about 2043. You would have to actually go all the way up to about 16.4% to have the trust fund be solvent until about 2085. And even then, eventually it will go negative, because benefits continue to rise.

Knowledge@Wharton: If you took some of the other measures you were talking about earlier, and then also did this for the interim period to get over the hump it sounds like that might work?

Smetters: Yes. Our policy simulator allows for 4,096 combinations — and so, lots of different combinations that people can try. And they get instant feedback on that. It’s a very sophisticated model, but we’ve used advances in cloud computing to pre-calculate everything for you, so you get instant feedback. So there are lots of combinations, even ones I haven’t had time to explore yet, that could potentially work here.

Knowledge@Wharton: We’ll get to the model that you talked about, that you helped develop, soon. As a matter of fact, that’s going to be something that anyone can go on the web and use, is that right?

Smetters: Everybody can use it. Policymakers, we’ve had a lot of conversations with them. They’re excited about it. They want access to tools that they can use while they’re actually writing legislation. But the average person can use it, and we’ve been showing it to various professors around the country, and they want to use it for their classrooms, just to give students tools where they can start to see these tradeoffs.

Knowledge@Wharton: The whole idea of immigration, illegal immigration — undocumented workers — has been a big part of this presidential primary process, and now the presidential race. So the $64,000 question is, will providing a path to legalization for unauthorized workers, immigrants, in the U.S. — will that improve the economy, or does it hurt the economy by hurting some lower-wage workers?

“We calculate that deportation actually will substantially lower GDP and lower jobs.”

Smetters: It turns out, and this is one of the surprises that we found, and one of the advantages of having this really detailed model — is that the impact on the economy from legalization is basically a wash. Here’s what’s going on. When it comes to jobs, legalization actually slightly reduces the number of jobs, and the number of people working. The reason why is that when you’re illegal, you pretty much have to be working, because you don’t have access to unemployment insurance, you don’t have access to going to school or college.

Knowledge@Wharton: You probably came here to work to begin with.

Smetters: You came here to work. And you pretty much have to work to eat. Once we have legalization, and we’re able to track in the census-level data the differences between undocumented and legal, what we see is legal immigrants have a little bit of a lower labor force attachment rate. The reason why is because they can spend more time looking for appropriate work. They qualify for unemployment insurance. And they also can get a college degree and up their skills. So that’s the reason why the effect on GDP is basically a wash — in particular, less employment, but more skills. And it has almost no impact.

Knowledge@Wharton: So the implication is that deporting illegal immigrant workers isn’t going to help the economy?

Smetters: Yes. In fact, some people on the other side say, “Well, deportation should be good for the economy,” and that turns out not to be true, either. In fact, we calculate that deportation actually will substantially lower GDP and lower jobs. The intuition behind that is that native workers, native-born workers — their labor force participation rate cannot possibly increase enough to offset the loss of jobs. And so the impact on GDP is actually fairly negative.

Knowledge@Wharton: Another question is, can increasing the proportion of authorized immigrants with college degrees, can that help our economy? A lot of kids come here from other countries, they get their degrees, and they’re not able to stay.

Smetters: Right. It’s really critical, as you just noted, to distinguish between — these are legal immigrants, as opposed to illegal immigrants. We have about 800,000 legal immigrants per year. About 35% of them have a college degree. So a lot of the proposals have been, “Let’s keep the 800,000 the same number, but shift that 35% upward.” Maybe make it half, 55%, and so forth. To our surprise, and this is the value of having these detailed models, it actually has a positive impact by increasing the number of college-educated immigrants. It has a positive impact, but it’s not that big.

If you dig into the numbers, what you see is that first, immigration itself, legal immigration as a total, is only about 1% of the U.S. population. Then when you go from 35% to 55%, you’re actually increasing the college-educated workforce in the United States by only about 0.5%. So the impact is positive, but it’s not a huge impact. Instead, what you ultimately need to do is to increase not just the shift, the change in composition — but actually increase the number of H-1Bs [a non-immigrant visa allowing employers to hire foreign workers temporarily for specialty occupations] without a change in the number of unskilled immigrants. That actually has a much bigger impact.

“Increased immigration is actually a slight improvement for Social Security and Medicare … we have more young people relative to old people….”

Knowledge@Wharton: What about lifting the total number of immigrants above 800,000?

Smetters: More immigrants basically mean more jobs and more GDP. In fact, what we find is that if we actually increase legal immigration from 800,000 by an additional 400,000 — by roughly 50% — by 2050 we’ll have a GDP that’s about 15% larger than what we see today.

Knowledge@Wharton: What is that annually? How much would GDP go up per year if you increased by that number?

Smetters: Over $1 trillion. We’re talking about a pretty significant number in current dollars. I would have to do the exact calculations, but it’s well over $1 trillion.

Knowledge@Wharton: Okay. What effect would an increase in immigration have on Social Security and Medicare?

Smetters: That used to be a very controversial issue. There was always this concern that immigrants are coming in right before they qualify for Social Security and Medicare. A lot of that debate was not correct, and a lot of changes have been made in the laws in the meantime. Nowadays, increased immigration is actually a slight improvement for Social Security and Medicare. And the reason why is immigrants are coming in during their working years, typically younger. As a result, what’s called the old-age dependency ratio is improved. In particular, we have more young people relative to old people, and that helps improve the finances. It’s not huge, but nonetheless it is positive.


Citing Knowledge@Wharton


For Personal use:

Please use the following citations to quote for personal use:


"Social Security and Immigration: A New Way to Model the Impact." Knowledge@Wharton. The Wharton School, University of Pennsylvania, 29 June, 2016. Web. 24 September, 2017 <>


Social Security and Immigration: A New Way to Model the Impact. Knowledge@Wharton (2016, June 29). Retrieved from


"Social Security and Immigration: A New Way to Model the Impact" Knowledge@Wharton, June 29, 2016,
accessed September 24, 2017.

For Educational/Business use:

Please contact us for repurposing articles, podcasts, or videos using our content licensing contact form.