Hemorrhaging Beneath Headlines of 1% Growth

OPT did not mask massive international enrollment declines, but a K-shaped enrollment cycle might be.

“It is a riddle, wrapped in a mystery, inside an enigma.” – Winston Churchill

Churchill’s famous phrase perfectly captures the contradictory signals we’re seeing in this fall’s international enrollment data in the United States.

SEVIS reports international student numbers grew 1% from September 2024 to September 2025 (1,222,247 → 1,231,470). Yet when I open my news and social media feeds, they’re full of reports about enrollment declines—some catastrophic.

So what gives?

Both can be right. Both are right. We need multiple data points to triangulate reality.

I’ve been warning for years about underlying structural fragility in the system that excessive focus on aggregate data obscures.

The reality most of us have seen coming for a while:

The year-to-year comparison of total enrollment masks uneven impacts and fragilities in the system. This is the jagged frontier of international enrollment. And the terrain is likely to become even more uneven.

Is OPT the X-Factor Masking Real Declines?

The “statistical cognitive dissonance” of 1% growth in SEVIS data has led some people (including me!) to speculate that perhaps OPT masked real declines of 10-15%. After all, we all lived through the “visa pause” and volatile policy environment and have been bracing for the worst.

Let’s test that hypothesis and start with what we know:

• SEVIS total (Sept 2024): 1,222,247 (all degree levels plus language training)

• SEVIS total (Sept 2025): 1,231,470 (+9,223, +1%)

Note: SEVIS counts all currently enrolled students plus recent graduates working on OPT—the work authorization that lets international graduates stay and work in the U.S. for one to three years.

Now let’s consider that OPT has grown substantially.

• OPT participation surged 22% year-over-year (2022/23: 198,793 → 2023/24: 242,782) in the last Open Doors report (and another 12% according to the 2024 Snapshot Report to about 2024/25: 271,916)

• Over the past 10 years, OPT has more than doubled (2014/15: 120,287 → 2024/25: 271,916)

Let’s test OPT as X-Factor hypothesis:

SEVIS Total - OPT Participants = Enrolled Students

SEVIS doesn’t disaggregate enrolled students from OPT. But we can plug in different scenarios based on past data and trends.

Here are three possibilities to get to 1% overall growth in SEVIS and what each would imply about enrollment growth or decline once OPT is taken into consideration:

• -15% OPT decline (231K) → +5% enrollment growth

• 0% OPT change (272K) → +1% enrollment growth

• +15% OPT growth (313K) → −3% enrollment decline

To fully “mask” real enrollment declines of 10-15%, OPT would need to have grown significantly.

• +40% OPT growth (381K) → −10% enrollment decline

• +55% OPT growth (421K) → −15% enrollment decline

It’s possible, but unlikely OPT increased 40-55% in a single year, so the “OPT masked massive declines” hypothesis likely will not hold. We will see. And I’ll be the first to admit it if I got this wrong.

Hemorrhaging Beneath the +1% Headline Number

The more likely explanation:

In statistical terms, +1% SEVIS growth is the “average” that likely reflects some buffering from continued growth in OPT, masking single-digit enrollment declines.

Underneath the 1% headline number, there’s likely significant variation in institutional impacts—ranging from modest growth to catastrophic declines.

And, remember, even if there were “only” single-digit enrollment declines (e.g., -6%), masked by growth in OPT, those declines represent a double-digit downward shift (e.g., -12%) in a sector used to single-digit annual growth (e.g., +6%) with more significant enrollment declines at institutions with the greatest risk exposure.

The distribution likely isn’t normal—it’s potentially wildly bifurcated, with small privates, regional publics, and other universities. Here are a few examples I was able to gather from news and institutional sources:

• Northwest Missouri State: −78% (557 → 125 students)

• Saint Louis University: −45% new graduate students (1,400 → 300)

• DePaul University: −30% (755 fewer students)

• Central Michigan: −23% (1,659 → 1,210)

• University at Buffalo: −15% (750–1,000 fewer graduate students)

Even some R1 flagship universities, like the University of Wisconsin–Madison, are facing extreme impact, dropping 30% year-over-year.

Meanwhile universities like Rutgers and University of Illinois–Urbana-Champaign have welcomed record numbers of international students (for this year, at least).

This is textbook Simpson’s Paradox—when aggregate trends mask or even reverse subgroup realities.

The forest looks stable while large numbers of trees are dying.

My sense is this:

If your institution’s numbers are flat or up, you’re staying quiet, seeking to not draw too much attention to your institution or “record” numbers of international students.

Meanwhile, institutions facing the steepest declines are making them known to sound the alarm for local, state, and national policymaker’s who might be concerned about potential impacts on the local and state economy.

This is the “signal-to-noise effect”—the signal is there but drowned out by louder but less representative stories.

Honestly, I’m surprised no major media outlet has done a deep dive on the SEVIS data. Hopefully reporters are digging behind the scenes, but the relative silence on this important dataset has been notable.

Risk Factors: What Institutions Are Most Exposed?

Here are some “risk factors” I anticipate may be behind institutions that have been hit hardest. To be clear, we won’t be able to test these factors until IIE releases its Fall Snapshot, but they’re worth considering as we took for patterns in the trends. (And feel free to add to this list in the comments.)

The best predictor isn’t a single factor, but the stacking of vulnerabilities:

1. Program Mix (Short-Cycle vs. Long-Cycle)

Heavy reliance on one-year master’s programs = structural fragility. These programs are most sensitive to visa timing, policy uncertainty, and OPT credibility. Institutions with balanced portfolios (including doctoral and undergraduate programs) show more resilience. Global demand for US master’s degrees has plunged by 60%.

2. Geographic/HALO Factors

Institutions in expensive metros without affordable housing or strong job markets see more attrition. “Value states”—Indiana, Texas, Florida, North Carolina—absorb displaced demand. Location matters more than ever when students are optimizing for total cost of ownership and post-graduation employment. HALO Factors = Housing, affordability, labor markets, and openness.

3. Source-Country Concentration

Institutions that leaned heavily on China, India, South Korea, or Japan as feeder markets faced bigger shocks, because those countries showed large declines. High concentration = high risk. Diversified portfolios = greater stability. Resilient institutions were able to diversify their recruitment strategy into emerging source countries. Those suffering the greatest losses likely remained dependent on traditional pipelines that suddenly dried up.

4. Admissions Flexibility / Recruitment Diversification

Institutions that hedged risk by over-admitting, recruiting more broadly geographically, or building relationships in emerging markets were better positioned when traditional pipelines contracted. Agility matters. Institutions that could pivot quickly—adjusting recruitment strategies, offering conditional admits, extending deadlines—weathered the storm better than those locked into rigid processes.

5. Brand Strength & Resource Capacity

Elite universities with strong brand pull (MIT, Stanford, Columbia, Michigan, UT Austin) can offset visa frictions through demand elasticity and generous financial aid. They have pricing power and reputational insurance. Mid-tier institutions without distinctive value propositions are less able to compete, especially under price pressure from international competitors offering similar quality at lower cost.

A K-Shaped Future of International Enrollment?

The SEVIS 1% growth offers some relief that perhaps things weren’t as bad some of us feared. But any relief should be temporary and only if we ignore the structural headwinds that could put international enrollment in reverse.

2025 is likely a harbinger of what’s ahead. It’s important to take in all the data and not ignore data that may conflict with our view of what’s happening.

Universities with globally recognized brands, generous financial aid, and strong OPT-to-H1B pipelines likely will remain resilient. MIT, Stanford, top publics, and specialized institutions can weather policy shocks because their value proposition—elite credentials plus U.S. career access—remains compelling. They likely used a mix of the five strategies (and may expand online programs to make up for any geographic limitations).

“The Death Spiral”

Meanwhile, it appears some universities face the risk of a “death spiral” in what might become an Age of Conquest if the data support a K-shaped patterns this cycle and if they continue moving forward.

The math gets brutal fast.

As international enrollment declines → institutions cut programs and services → campus experience degrades → becomes less attractive to all students → enrollment declines further → death spiral accelerates.

A 40% decline in international graduate students doesn’t mean “tighten belts 40%”—it sometimes means “we can’t run these programs at all.” When graduate programs close, reputation declines. The unraveling compounds.

Looking Ahead: Three Critical Questions

Distributed Progress is not about “today’s news” — it’s about the future we want to create for tomorrow.

My time horizon is not reacting to today’s policy shock but tracing the long-term trends reshaping internationalization in the US and worldwide.

So what are scenarios we need to prepare for in the next year or so:

What If OPT Growth Plateaus?

OPT has been expanding as a compounding rate — 22% in the last year-over-year — bolstering headlines about the total number of international students. If the OPT prop falters—whether through slowed growth, declines, or stronger competitors—that’s when policymakers will see what universities have been experiencing all along.

What If OPT Is Eliminated?

If OPT were ended—as has been proposed by Joseph Edlow, the current Director of U.S. Citizenship and Immigration Services (USCIS)—headline growth would collapse into visible decline, exposing the fragility of the master’s pipeline. This week, U.S. Sen. Charles Grassley called on the Department of Homeland Security to stop approving work authorization for OPT. The value proposition for international graduate education depends heavily on OPT. If it is eliminated so is one of the primary reasons students choose U.S. programs over alternative countries.

How Much is the Policy Uncertainty Tax?

Even temporary visa restrictions create long-tail reputation effects. International students make decisions 12–18 months in advance. Every policy disruption—from H-1B to uncertainty about the impacts of duration of status (D/S) rules—introduce friction into multi-year planning cycles. Students and families increasingly treat U.S. higher education as higher risk, lower certainty. That perception doesn’t reverse quickly, even when policies stabilize, after student flows have redirected elsewhere.

The Test of First-Rate Intelligence

I’m a realist. I update my priors when evidence contradicts my assumptions.

• The SEVIS 1% growth tells us something important, but not the whole story.

• The headlines about massive enrollment declines at some universities, even flagships, tell us something important, but not the whole story.

My sense is that when this year’s Open Doors data is released in November 2026 (a long time to wait!), we’re likely seeing the most uneven enrollment cycle in recent history. Time will tell. I may be wrong. But that’s how I’m making sense of these “mixed signals” at the moment.

Regardless, the real story has always existed under the headline numbers, even if they are not what the media reports. The “signal” in the “noise” is often the variation, the divergence beneath the stable-looking year-over-year growth.

F. Scott Fitzgerald said:

“The test of a first-rate intelligence is the ability to hold two opposed ideas in the mind at the same time, and still retain the ability to function.”

The “mixed signals” are a test of our field’s intelligence. To learn and adapt. We can triangulate “conflicting” data as we navigate new realities. Headlines won’t do.