OPINION AND ANALYSIS:
Below is our conversation with ChatGPT on 5/5/26 regarding whether or not we need immigration.
IMMIGRATION & SOCIAL SECURITY INSOLVENCY (BIRTH RATES)
Rx. Is it not true that social security cannot be put on a realistic sustainable plan as long as birth and fertility rates are below 2.1 per couple. In answering assume that (i) there is no immigration; and (ii) the phrase “ realistic sustainable plan” is a plan that pays out benefits at least at a rate that are at least 90 percent that of the present value of current benefit levels with a start date of benefits being no later than the age of 70.
Show moreShow less
David, I’ll treat this as a demographics-and-finance question, not a political slogan. I’ll check the current Social Security trustee framework and then answer your exact hypothetical: no immigration, sub-replacement fertility, 90% benefits, start no later than 70.
Thought for 10s
David, yes — in your no-immigration hypothetical, I think that statement is basically true as a practical matter, with one lawyerly caveat.
If fertility stays below replacement and no immigration fills the worker pipeline, Social Security’s pay-as-you-go structure becomes increasingly strained because fewer workers support more retirees. The 2025 Trustees already assume fertility eventually rises only to 1.9, plus substantial immigration, and even then OASDI has a 3.82% of taxable payroll 75-year deficit; after depletion, only 81% of scheduled benefits would be payable in 2034, declining to 72% by 2099.
Your added standard — at least 90% of current benefit value and benefits starting no later than 70 — makes the problem much harder. Under current projections, automatic payable benefits would fall below that 90% floor even with immigration. Remove immigration, and the worker-to-beneficiary ratio worsens further.
The caveat: Congress could technically “solve” it by very large tax increases, general Treasury transfers, wealth taxes, VAT-style taxes, or major benefit redistribution. So it is not mathematically impossible.
But if by “realistic sustainable plan” you mean a Social Security-like system funded mainly by payroll taxes while preserving at least 90% benefits by age 70, then sub-2.1 fertility plus zero immigration makes that unrealistic over the long run. In plain English: without enough babies or immigrants, the system eventually runs short of workers. The money tree needs leaves.
Rx. I am asking this question because to me it’s profoundly relevant and it’s answer is very important to resolving a huge difference between Republicans (Trump/MAGA) and what some Democrats think. Namely, when (i) it comes to making social security solvent you really can’t do it (even if you delay benefits) even if Congress raises payroll or other taxes to fund it, without immigration if birth rates are below 2.1
Show moreShow less
David, I’ll tighten the answer around the exact political/economic proposition: whether tax increases alone can make Social Security solvent without immigration when fertility is below replacement. I’m going to separate “accounting solvency” from “real-world sustainability,” because that distinction matters here.
Thought for 6s
David, I would phrase it this way:
You can make Social Security “actuarially solvent” with taxes alone, even with low fertility and no immigration — but that may not be a realistic sustainable plan in the broader economic sense.
The key distinction:
- Accounting solvency:
Congress can raise taxes enough to pay promised benefits. CBO estimated that restoring 75-year solvency could be done by an immediate permanent payroll-tax increase of about 4.3 percentage points, from 12.4% to 16.7%, under its 2024 assumptions. - Real-world sustainability:
If fertility remains below replacement and immigration is zero, the worker-to-retiree ratio keeps deteriorating. That means each worker must support a larger retired population. At some point, the tax burden needed to preserve 90% benefits by age 70 may become economically and politically unrealistic.
So your strongest defensible point is:
Without immigration, and with birth rates below replacement, Social Security cannot be kept both generous and politically/economically realistic merely by delaying retirement or modest tax increases.
But I would avoid saying “cannot be done even if Congress raises payroll or other taxes” as an absolute. A government can always impose very high taxes or transfer general revenues. The better argument is that taxes alone become increasingly unrealistic because demographics are doing arithmetic against the system.
The MAGA-vs-Democratic divide, stated plainly:
If America will not have enough children and also refuses enough immigration, then preserving Social Security at near-current benefit levels requires much higher taxes, much later retirements, or much lower benefits. AI may raise productivity, but AI does not by itself create payroll-tax-paying workers unless its gains are captured and taxed.
That last sentence is the heart of it.
Is it not true. That the longer we wait to get more workers into the social security system the more difficult it will be to make social security solvent?
Show moreShow less
Yes, David — that is generally true.
Social Security is heavily dependent on the ratio of workers paying payroll taxes to retirees receiving benefits. If the country delays increasing the number of workers — whether through higher birth rates, immigration, or higher labor-force participation — then several things happen simultaneously:
- More Baby Boomers retire and collect benefits.
- Fewer younger workers replace them.
- The trust fund gets depleted further.
- The remaining time for new workers to contribute shrinks.
That last point is crucial.
A 25-year-old immigrant worker today might pay payroll taxes for 40 years before full retirement age. But if the country waits 15 years before adding that worker, the system loses 15 years of contributions that can never be recovered.
Similarly, babies eventually become workers — but not for roughly 20+ years. So declining fertility has a delayed but powerful effect.
Economically, this is sometimes described as a “dependency ratio” problem:
- fewer workers supporting more retirees.
And because Social Security is largely a pay-as-you-go system rather than a fully invested pension fund, timing matters enormously.
So your proposition can fairly be stated this way:
The longer the United States delays increasing the number of payroll-tax-paying workers, the more difficult and expensive it becomes to maintain Social Security benefits at near-current levels.
That proposition is strongly supported by mainstream demographic and actuarial analysis.
So, if we wait until 2032 or 2033 or 2034 to try to make social security solvent the amount of tax increase and/or the reduction in benefits required may be impossible to achieve?
Show moreShow less
David, yes — that is a fair economic and political concern, though “impossible” is probably too absolute a word if you are speaking analytically rather than rhetorically.
The basic actuarial principle is correct:
The longer corrective action is delayed, the larger and more abrupt the required tax increases or benefit reductions become.
The reason is simple:
- once the trust fund reserves are exhausted, the system loses the cushion that temporarily bridges the gap between payroll taxes coming in and benefits going out;
- every year of delay means fewer years over which to spread the adjustment;
- meanwhile the retired population continues to grow relative to the working population.
The Social Security Administration Trustees have repeatedly stated that earlier reforms permit:
- smaller tax increases,
- smaller benefit reductions,
- and more gradual phase-ins.
By contrast, waiting until around projected depletion years (currently about 2034 for combined trust fund depletion under many projections) would likely require:
- sharp payroll-tax increases,
- substantial general revenue transfers,
- significant benefit cuts,
- higher retirement ages,
- or some combination of those measures.
So your core point is economically sound:
Delay substantially increases the magnitude and political difficulty of the required fixes.
Now, could Congress technically still fix it in 2034?
Yes.
Congress has enormous taxing and borrowing power.
But your broader point is that the required measures may become so politically unpopular or economically disruptive that achieving consensus becomes extraordinarily difficult. That concern is widely shared by actuaries and economists across the political spectrum.
One additional point relevant to your immigration argument:
workers added earlier are disproportionately valuable because they contribute payroll taxes for many years before drawing benefits. Delayed action loses those contribution years permanently.
Rx. My point is that as a practical matter its impossible to make social security solvent and then keep it solvent in the long term without enough immigration to make up for the shortfall between the todays fertility/birth rates and fertility birth rates of 2.1. Because once the amount of immigration does not make up that birth/fertility rate shorfall, social security will immediately start to develop an insolvency problem again
Show moreShow less
David, your argument is logically strong, but I would tighten it slightly to avoid overstating it.
This version is closer to what mainstream demographers and actuaries would likely accept:
As a practical long-term matter, if fertility remains materially below replacement level, then absent sufficient immigration or dramatically higher worker productivity, Social Security will repeatedly drift back toward insolvency unless taxes are continually increased or benefits continually reduced.
That is the key point.
The reason is structural:
Social Security depends primarily on current workers funding current retirees.
If:
- births stay below replacement,
- immigration does not offset the shortfall,
- and retirees live longer,
then over time the worker-to-beneficiary ratio keeps deteriorating.
Historically, the ratio was roughly:
- about 5 workers per beneficiary in 1960,
- around 2.7 today,
- projected near 2.2 in coming decades.
At some point, maintaining current-style benefits requires either:
- much higher taxes per worker,
- much lower benefits per retiree,
- much later retirement ages,
- substantial outside revenue transfers,
- more workers (immigration or higher birth rates),
- or dramatically higher productivity.
So the strongest version of your proposition is probably:
Immigration is not merely a cultural or humanitarian issue; under current demographic trends it functions as an important economic mechanism for sustaining programs like Social Security.
Now, where critics of your position would push back is here:
they would argue AI/productivity growth could offset the declining worker ratio.
That is not a frivolous argument. If each worker becomes vastly more productive, then fewer workers can theoretically support more retirees.
But your counterpoint is also legitimate:
Social Security is funded mainly through payroll taxes on workers. If AI wealth accrues mainly to capital owners rather than wage earners, then productivity gains alone do not automatically solve Social Security financing.
That is really the heart of the modern debate.
My poin tis that all the AI productivity arguements are totally speculative. We have no history to base estimates on. We have no idea, how much more productive and/or when that productivity increase will occur. And frankly, it’s just like Musk’s promise of self driving cars, some of it is sheer salesmanship
Show moreShow less
David, that is a reasonable critique, and many economists would agree with at least part of it.
Your central point is essentially this:
Demographic decline is a present, measurable actuarial problem, whereas massive AI-driven productivity gains sufficient to offset it are largely projections and assumptions about the future.
That distinction matters.
We already know with fairly high confidence:
- current fertility rates,
- retirement trends,
- life expectancy ranges,
- current worker-to-beneficiary ratios,
- and projected Social Security cash-flow gaps.
Those are actuarial inputs grounded in existing demographic data.
By contrast, claims that AI will:
- massively increase productivity,
- preserve employment,
- increase wages broadly,
- generate taxable payroll income,
- and do so soon enough to stabilize Social Security,
are much more uncertain.
Your skepticism becomes even stronger because:
- prior technological revolutions often took decades to diffuse through the economy;
- productivity gains do not always translate into wage gains;
- wealth from automation may concentrate in capital ownership rather than payroll income;
- and some AI may reduce payroll-tax-paying jobs rather than increase them.
So your argument is not “AI cannot help.”
Your argument is:
It is risky public policy to assume speculative future AI gains will solve a mathematically predictable demographic problem.
That is a defensible position.
And your Musk/self-driving analogy makes the rhetorical point effectively:
promises about transformative technology are often overly optimistic regarding timing, scale, and reliability.
One thing I would add, though, for balance:
some economists would respond that America has repeatedly underestimated technological productivity gains over long periods — railroads, electrification, computers, the internet, etc.
So the disagreement is less about whether AI could help, and more about:
- how much,
- how fast,
- who benefits economically,
- and whether government can capture enough of those gains to finance entitlement systems.