Results

Main DiD Estimates

Table 2 in the paper presents the main difference-in-differences estimates for all three outcomes.

(1) New Entry (2) Enrollment (3) Graduation
Post × Treatment −13.315*** −8.199*** −7.542***
(1.806) (3.171) (1.290)
Post 4.586** 7.416*** 2.099*
(1.853) (2.409) (1.195)
Treatment 3.320*** 3.516 0.810
(0.980) (3.645) (1.087)
Controls Yes Yes Yes
R² 0.08 0.11 0.09
Observations 50,670 50,670 50,670
Pre-COVID mean 83.195 322.339 34.922
Effect as % of pre-COVID mean 16% 2.54% 21.55%

Source: ANUIES. Standard errors in parentheses, clustered at the municipality level. Controls include a dummy for delivery format, a dummy for whether the university is public, and a dummy for whether the university is top-20. * p<0.10, ** p<0.05, *** p<0.01.

The headline result. The pandemic reduced new entry by 16%, graduation by 21.5%, and enrollment by 2.5% relative to pre-pandemic levels. Graduation was hit hardest. The relatively small enrollment effect is consistent with students remaining formally registered while reducing their course load or delaying completion.

Heterogeneous Effects by Institution Type

Table 4 in the paper estimates the DDD specification for three dimensions of institutional heterogeneity.

Public vs. Private Universities

New entry
+1.04%
Differential effect for public universities. Null — no significant difference in new entry between public and private.
Enrollment
+1.41%
Differential effect for public universities. Null — no significant difference in enrollment.
Graduation
−27.2%
Public universities suffered a significantly larger graduation decline, relative to the public pre-COVID mean of 74.3 graduates per program.

Public institutions performed significantly worse than private ones in graduation outcomes, but not in new entry or enrollment. The pre-pandemic mean for public graduation was 74.3 versus 17.8 for private institutions, making the public sector’s higher absolute losses especially consequential.

Top-20 vs. Non-Top-20 Universities

New entry
+7.99%
Top-20 universities gained new students relative to non-top-20. Significant at p<0.01.
Enrollment
+2.21%
Top-20 universities also gained in enrollment relative to non-top-20. Significant at p<0.05.
Graduation
−20.4%
Top-20 universities experienced a larger graduation decline relative to their pre-COVID mean of 128.6 graduates per program. Marginally significant (p<0.10).

Elite universities attracted more new students during the pandemic, consistent with a flight-to-quality response: students anticipated that well-resourced institutions would handle the disruption better and shifted preferences accordingly. The best schools in Mexico tend to attract students with higher incomes and resources, so this pattern is also consistent with the pandemic disproportionately affecting lower-income students.

Synchronous vs. Asynchronous Programs

In-person programs showed no significant difference from asynchronous (already online) programs in new entry or enrollment. However, synchronous programs experienced a worse graduation outcome: the differential graduation effect was −11.45% relative to the synchronous pre-COVID mean (41.5 graduates per program). This is consistent with in-person programs facing larger administrative hurdles for degree certification during campus closures.

A surprising null result. One might expect programs already designed for remote delivery to have a natural advantage during the pandemic. The data show that having an online format did not protect institutions from new-entry or enrollment shocks. The disruption appears to be driven by the income and health shock component of the pandemic, rather than by the transition to online delivery per se.

Effects by Gender

Table 5 presents the DiD estimates separately for men and women, overall and by STEM vs. non-STEM fields.

New Entry Enrollment Graduation
Men (overall) −17.18%*** −3.52%*** −21.03%***
Women (overall) −14.88%*** −1.58% −22.08%***
Men — STEM −13.94%*** −2.68%*** −23.02%***
Women — STEM −10.55%*** −1.04% −24.02%***
Men — Non-STEM −20.20%*** −4.39%*** −19.01%***
Women — Non-STEM −17.00%*** −1.85% −20.92%***

All estimates as % of pre-COVID mean. *** p<0.01. Enrollment effects for women are not statistically significant in any specification. Source: ANUIES, Table 5.

Key gender patterns. New entry and graduation declines are statistically significant for both men and women across all specifications. The enrollment decline is significant for men but not for women in any category — overall, STEM, or non-STEM — suggesting women remained formally registered at similar rates to the pre-pandemic period even as their new entry and graduation were disrupted. Men experienced larger new entry and enrollment declines; women experienced a slightly larger graduation decline, consistent with evidence that female students faced greater difficulty completing their degrees once the pandemic hit. Within STEM fields, graduation declines reach approximately 23-24% for both genders, above the overall average.

Effects by Area of Study

Table 6 presents the DiD estimates for all ten areas of study. Graduation rates and new entry showed the most variation across fields.

Area of Study New Entry (% change) Significant?
Education −25.25% ***
Services −24.54% ***
Social Sciences −18.00% ***
Business −17.95% ***
Engineering −13.45% ***
Arts & Humanities −13.35% ***
Health −12.22% ***
Information Technology −12.02% ***
Agronomy & Veterinary −9.33% **
Sciences −5.99% (not significant)

Education-related majors suffered the highest new-entry decline (25.25%). Sciences showed no significant effect. Social sciences and services were also severely affected (18% and 24.5%).

Area of Study Graduation (% change) Significant?
Sciences −37.5% **
Agronomy & Veterinary −27.35% **
Health −24.32% ***
Social Sciences −23.9% ***
Engineering −22.1% ***
Services −22.23% ***
Business −18.9% ***
Information Technology −20.09% ***
Education −13.87% **
Arts & Humanities −11.67% **

Sciences experienced the most severe graduation collapse (37.5%), followed by agronomy and health. Arts & humanities and education had the smallest — but still significant — graduation declines.

Two fields of particular policy concern. Sciences saw a 37.5% collapse in graduation — the single largest effect in the paper — with implications for STEM pipeline development. Education-related majors saw a 25% drop in new entry, raising the prospect of future teacher shortages if this trend persists or is not reversed. As Aucejo et al. (2020) reported, approximately 12% of students considered changing majors due to the pandemic, suggesting some of this new-entry decline may reflect permanent preference shifts.

Robustness

The treatment window is shifted one year earlier: treatment = academic years 2018-2019 and 2019-2020, control = 2016-2017 and 2017-2018. If the parallel trends assumption holds, we should find null effects. As expected, the DD effects were generally insignificant, except for enrollment effects in education, agronomy, and arts and humanities-related majors and new entry effects in engineering, arts and humanities, and services majors, which showed significant differences.

New Entry Enrollment Graduation
Post × Treatment (placebo) 0.439 3.108 −0.991
(0.962) (4.251) (0.769)
R² 0.07 0.11 0.10
Observations 48,194 48,194 48,194

Panel B of Table 3 presents p-values and FDR-corrected q-values. None of the statistically significant results becomes insignificant after the multiple-testing correction.

New Entry Enrollment Graduation
Post × Treatment −13.315*** −8.199*** −7.542***
p-value (4.74e-13) (0.009) (7.72e-09)
q-value [0.001] [0.005] [0.001]
Observations 50,670 50,670 50,670

Panel C of Table 3 presents Oster bounds based on an expected R² value. Bounds that exclude zero indicate robustness to omitted variable bias. All three bounds exclude zero.

New Entry Enrollment Graduation
Oster bounds [−23.062, −11.206] [−9.028, −4.386] [−10.024, −7.011]
Observations 50,670 50,670 50,670

Panel D of Table 3 re-estimates the main DiD ten times, each time omitting one area of study. All estimates remain highly significant and close to the main specification, confirming that results are not driven by any single field.

Specification New Entry Enrollment Graduation
Overall (baseline) −13.315*** −8.199*** −7.542***
Leaving out sciences −13.442*** −8.511*** −7.342***
Leaving out social sciences −12.824*** −7.514** −7.270***
Leaving out education −13.495*** −9.335*** −8.081***
Leaving out engineering −12.628*** −6.838** −6.755***
Leaving out health −12.921*** −7.701** −6.887***
Leaving out inf. technology −14.254*** −9.423*** −8.008***
Leaving out business −12.528*** −7.293*** −7.734***
Leaving out agronomy & vet. −13.421*** −8.030** −7.500***
Leaving out arts & hum. −13.941*** −8.752** −8.046***
Leaving out services −13.502*** −8.369** −7.805***