A Decade of Centralized School Choice Admission in Chile: Achievements and Challenges

Christopher A. Neilson

Paper · Working Paper · Working paper (Draft: May 2026) · 2026

A Decade of Centralized School Choice Admission in Chile: Achievements and Challenges

Abstract

This paper reviews the first decade of Chile's Sistema de Admisión Escolar (SAE), a national school assignment institution created under the 2015 Inclusion Law to replace fragmented school-by-school admissions with a coordinated platform for the publicly funded sector. Around the world, access to publicly funded education is rarely allocated only through prices. Societies therefore need rules for rationing scarce seats: how preferences count, who receives priority, and how principles such as equal opportunity, sibling priority, and vulnerability should be implemented when demand exceeds supply. Chile's reform is a useful case because it did not simply adopt a matching algorithm. It built a national institution around three pillars: reducing strategic incentives across the design, coordinating public and private-subsidized schools through one platform, and making the algorithm, priorities, and data transparent enough to support evaluation and public scrutiny. The paper organizes the decade around four questions: what was implemented, how Chile learned while scaling, what happened after a decade, and what remains unfinished. The first answer is that SAE was more than Deferred Acceptance. The design combined a low-strategy assignment rule with unrestricted preference lists, legally defined priorities, lottery-based tie-breaking, sibling and family rules, fallback provisions, complementary-round rules, national coverage, and public documentation. These features reduced many of the discretionary and strategic margins that characterized the pre-SAE system. They also created an administrative and research infrastructure that made the system observable as it scaled. The second answer is that implementation design mattered almost as much as mechanism design. SAE was rolled out regionally between 2016 and 2019, with national implementation consolidated by 2020. This staggered rollout created a research-policy loop: administrative data, surveys, experiments, quasi-experimental variation, and annual technical recommendations could be connected to family behavior inside the platform. Much was unknown ex ante. It was not clear whether families would trust a centralized assignment process, submit sufficiently broad lists, understand admission chances, or respond to information. Nor was it clear where local scarcity would bind, how sibling and fallback rules would work in practice, or how quickly the system could adapt. The central lesson from this learning process is that reducing strategic ranking incentives is necessary but not sufficient. Even under a low-strategy assignment mechanism, families still need to discover schools, compare quality and cost, understand admission chances, and realize when their current application portfolio exposes them to non-placement risk. The evidence reviewed in the paper shows that families often overestimate admission chances, underestimate feasible nearby options, and stop searching too early. These frictions are not merely informational details; they affect who can use the assignment system effectively. SAE's response was to develop a "smart platform" layer on top of the assignment mechanism. Risk warnings, personalized feedback reports, the School Explorer/MIME tools, WhatsApp nudges, and later Anótate en la Lista all use platform data to make options, risks, and consequences more visible to families. The evidence points to a consistent pattern: platform-embedded information can change application behavior. Families add options, reduce predicted non-placement risk, and in several studies move toward schools with higher measured value-added. New follow-up evidence is consistent with some of these gains translating into later learning outcomes, though the paper treats those results conservatively and distinguishes published evidence from new and companion working-paper results. The paper also adds three empirical contributions to the existing evidence. First, it reports an out-of-sample replication of the risk-warning regression discontinuity after the original QJE study window, covering SAE cycles from 2018 to 2023 and linking later cohorts to fourth-grade SIMCE outcomes. Second, it uses linked administrative records to quantify the new in-year aftermarket platform, Anótate en la Lista, showing that off-cycle applications are comparable in scale to the main round but still operate largely outside SAE's equity priorities. Third, it presents staggered event-study evidence on segregation, showing convergence in prioritario representation at oversubscribed schools, where the assignment mechanism actually binds. After a decade, SAE's aggregate assignment performance is strong but bounded. In 2025, 82.6 percent of applicants were assigned to one of their top three listed schools and 92.8 percent to a listed preference. These outcomes place Chile within the range of mature centralized systems. But the remaining 7.2 percent non-placement rate is geographically concentrated, and national averages hide local markets where desired seats are scarce. The evidence is consistent with local supply constraints shaping what SAE can deliver: demographic growth and migration increased demand in some areas, private-subsidized school entry slowed sharply, and some desirable high-value-added private-subsidized schools exited to the unsubsidized private sector. As a result, empty seats in some markets coexist with acute shortages in others. The equity effects are similarly bounded. SAE appears to have improved access at oversubscribed schools, where discretionary screening previously mattered most and where the allocation rule has real bite. However, aggregate segregation effects are harder to identify causally and depend on the measure. Residential stratification, the unsubsidized private sector, school entry and exit, and local supply constraints all operate outside the assignment algorithm. The right interpretation is therefore neither triumphalist nor dismissive: SAE delivered measurable gains in access at constrained, high-demand schools, while leaving broader segregation and quality-access problems for complementary policy. The unfinished agenda follows directly from these limits. First, assignment policy must be connected more explicitly to supply policy. A centralized mechanism can allocate scarce seats more transparently, but it cannot create desirable seats where too few exist. Second, the aftermarket needs stronger design. Anótate en la Lista made in-year demand visible at scale, but its first-come-first-served logic does not yet incorporate the equity priorities of the main round. Third, many algorithmic and priority-rule adjustments remain to be debated and evaluated, including sibling logic, fallback rules, tie-breaking, capacity provisions, and priority definitions. Fourth, communication with families remains a policy margin: families need clearer, timely, personalized information about risk, feasible options, sibling trade-offs, fallback consequences, and aftermarket rules. The broader message is that Chile's experience is best understood as institution plus learning. SAE built a transparent national assignment system, used staggered implementation to learn while scaling, and developed smart-platform tools to address the behavioral frictions revealed by that process. The next decade requires extending the same iterative discipline to the margins that have changed more slowly: supply, the aftermarket, algorithm and priority adjustments, communication with families, and public governance of complex assignment rules. Centralized assignment can make school access fairer, more transparent, and more empirically governable, but the mechanism is only one part of the institution. Its long-run value depends on whether the system can keep learning from its own operation and translate that learning into rules, tools, and policies that families can trust.

Citation & BibTeX

Christopher A. Neilson, "A Decade of Centralized School Choice Admission in Chile: Achievements and Challenges", Working paper (Draft: May 2026), 2026.

@article{ sae-decade_2026,
title = "A Decade of Centralized School Choice Admission in Chile: Achievements and Challenges",
author = "Neilson, Christopher A.",
journal = "Working paper (Draft: May 2026)",
year = "2026",
note = "wp"
,
abstract = "This paper reviews the first decade of Chile's Sistema de Admisión Escolar (SAE), a national school assignment institution created under the 2015 Inclusion Law to replace fragmented school-by-school admissions with a coordinated platform for the publicly funded sector. Around the world, access to publicly funded education is rarely allocated only through prices. Societies therefore need rules for rationing scarce seats: how preferences count, who receives priority, and how principles such as equal opportunity, sibling priority, and vulnerability should be implemented when demand exceeds supply. Chile's reform is a useful case because it did not simply adopt a matching algorithm. It built a national institution around three pillars: reducing strategic incentives across the design, coordinating public and private-subsidized schools through one platform, and making the algorithm, priorities, and data transparent enough to support evaluation and public scrutiny.The paper organizes the decade around four questions: what was implemented, how Chile learned while scaling, what happened after a decade, and what remains unfinished. The first answer is that SAE was more than Deferred Acceptance. The design combined a low-strategy assignment rule with unrestricted preference lists, legally defined priorities, lottery-based tie-breaking, sibling and family rules, fallback provisions, complementary-round rules, national coverage, and public documentation. These features reduced many of the discretionary and strategic margins that characterized the pre-SAE system. They also created an administrative and research infrastructure that made the system observable as it scaled.The second answer is that implementation design mattered almost as much as mechanism design. SAE was rolled out regionally between 2016 and 2019, with national implementation consolidated by 2020. This staggered rollout created a research-policy loop: administrative data, surveys, experiments, quasi-experimental variation, and annual technical recommendations could be connected to family behavior inside the platform. Much was unknown ex ante. It was not clear whether families would trust a centralized assignment process, submit sufficiently broad lists, understand admission chances, or respond to information. Nor was it clear where local scarcity would bind, how sibling and fallback rules would work in practice, or how quickly the system could adapt.The central lesson from this learning process is that reducing strategic ranking incentives is necessary but not sufficient. Even under a low-strategy assignment mechanism, families still need to discover schools, compare quality and cost, understand admission chances, and realize when their current application portfolio exposes them to non-placement risk. The evidence reviewed in the paper shows that families often overestimate admission chances, underestimate feasible nearby options, and stop searching too early. These frictions are not merely informational details; they affect who can use the assignment system effectively.SAE's response was to develop a \"smart platform\" layer on top of the assignment mechanism. Risk warnings, personalized feedback reports, the School Explorer/MIME tools, WhatsApp nudges, and later Anótate en la Lista all use platform data to make options, risks, and consequences more visible to families. The evidence points to a consistent pattern: platform-embedded information can change application behavior. Families add options, reduce predicted non-placement risk, and in several studies move toward schools with higher measured value-added. New follow-up evidence is consistent with some of these gains translating into later learning outcomes, though the paper treats those results conservatively and distinguishes published evidence from new and companion working-paper results.The paper also adds three empirical contributions to the existing evidence. First, it reports an out-of-sample replication of the risk-warning regression discontinuity after the original QJE study window, covering SAE cycles from 2018 to 2023 and linking later cohorts to fourth-grade SIMCE outcomes. Second, it uses linked administrative records to quantify the new in-year aftermarket platform, Anótate en la Lista, showing that off-cycle applications are comparable in scale to the main round but still operate largely outside SAE's equity priorities. Third, it presents staggered event-study evidence on segregation, showing convergence in prioritario representation at oversubscribed schools, where the assignment mechanism actually binds.After a decade, SAE's aggregate assignment performance is strong but bounded. In 2025, 82.6 percent of applicants were assigned to one of their top three listed schools and 92.8 percent to a listed preference. These outcomes place Chile within the range of mature centralized systems. But the remaining 7.2 percent non-placement rate is geographically concentrated, and national averages hide local markets where desired seats are scarce. The evidence is consistent with local supply constraints shaping what SAE can deliver: demographic growth and migration increased demand in some areas, private-subsidized school entry slowed sharply, and some desirable high-value-added private-subsidized schools exited to the unsubsidized private sector. As a result, empty seats in some markets coexist with acute shortages in others.The equity effects are similarly bounded. SAE appears to have improved access at oversubscribed schools, where discretionary screening previously mattered most and where the allocation rule has real bite. However, aggregate segregation effects are harder to identify causally and depend on the measure. Residential stratification, the unsubsidized private sector, school entry and exit, and local supply constraints all operate outside the assignment algorithm. The right interpretation is therefore neither triumphalist nor dismissive: SAE delivered measurable gains in access at constrained, high-demand schools, while leaving broader segregation and quality-access problems for complementary policy.The unfinished agenda follows directly from these limits. First, assignment policy must be connected more explicitly to supply policy. A centralized mechanism can allocate scarce seats more transparently, but it cannot create desirable seats where too few exist. Second, the aftermarket needs stronger design. Anótate en la Lista made in-year demand visible at scale, but its first-come-first-served logic does not yet incorporate the equity priorities of the main round. Third, many algorithmic and priority-rule adjustments remain to be debated and evaluated, including sibling logic, fallback rules, tie-breaking, capacity provisions, and priority definitions. Fourth, communication with families remains a policy margin: families need clearer, timely, personalized information about risk, feasible options, sibling trade-offs, fallback consequences, and aftermarket rules.The broader message is that Chile's experience is best understood as institution plus learning. SAE built a transparent national assignment system, used staggered implementation to learn while scaling, and developed smart-platform tools to address the behavioral frictions revealed by that process. The next decade requires extending the same iterative discipline to the margins that have changed more slowly: supply, the aftermarket, algorithm and priority adjustments, communication with families, and public governance of complex assignment rules. Centralized assignment can make school access fairer, more transparent, and more empirically governable, but the mechanism is only one part of the institution. Its long-run value depends on whether the system can keep learning from its own operation and translate that learning into rules, tools, and policies that families can trust."
,
url = "https://www.christopher-neilson.com//work/documents/SAE10/SAE10_paper_draft_may2026.pdf"
,
url_article = "https://www.christopher-neilson.com//work/documents/SAE10/SAE10_paper_draft_may2026.pdf"
}

Overview

This draft synthesizes the first decade of Chile’s national school assignment platform (SAE), connecting institution design, implementation evidence, and outcomes observed at scale. The paper frames SAE as both a matching mechanism and an evolving policy platform where operational data can be used to diagnose frictions and test improvements over time.

Key findings highlighted in the draft

Assignment outcomes after a decade: by 2025, 82.6% of applicants are assigned to one of their top three listed schools, and 92.8% to a listed preference.
Core mechanism lesson: strategy-proof ranking incentives help, but families still need better information and search support to choose feasible, high-value options.
Evidence on information tools: risk warnings and related platform interventions improve list composition and assignment safety, including in replicated analyses.
Persistent structural constraints: non-placement remains concentrated in local markets with binding seat scarcity and constrained supply adjustment.
Distributional change: event-study evidence points to improved prioritario representation at oversubscribed schools over time.
Aftermarket challenge: a large in-year market outside main-round priorities remains central for equity and governance.

Extended abstract

This paper reviews the first decade of Chile’s Sistema de Admisión Escolar (SAE), a national school assignment institution created under the 2015 Inclusion Law to replace fragmented school-by-school admissions with a coordinated platform for the publicly funded sector. Around the world, access to publicly funded education is rarely allocated only through prices. Societies therefore need rules for rationing scarce seats: how preferences count, who receives priority, and how principles such as equal opportunity, sibling priority, and vulnerability should be implemented when demand exceeds supply. Chile’s reform is a useful case because it did not simply adopt a matching algorithm. It built a national institution around three pillars: reducing strategic incentives across the design, coordinating public and private-subsidized schools through one platform, and making the algorithm, priorities, and data transparent enough to support evaluation and public scrutiny.

The paper organizes the decade around four questions: what was implemented, how Chile learned while scaling, what happened after a decade, and what remains unfinished. The first answer is that SAE was more than Deferred Acceptance. The design combined a low-strategy assignment rule with unrestricted preference lists, legally defined priorities, lottery-based tie-breaking, sibling and family rules, fallback provisions, complementary-round rules, national coverage, and public documentation. These features reduced many of the discretionary and strategic margins that characterized the pre-SAE system. They also created an administrative and research infrastructure that made the system observable as it scaled.

The second answer is that implementation design mattered almost as much as mechanism design. SAE was rolled out regionally between 2016 and 2019, with national implementation consolidated by 2020. This staggered rollout created a research-policy loop: administrative data, surveys, experiments, quasi-experimental variation, and annual technical recommendations could be connected to family behavior inside the platform. Much was unknown ex ante. It was not clear whether families would trust a centralized assignment process, submit sufficiently broad lists, understand admission chances, or respond to information. Nor was it clear where local scarcity would bind, how sibling and fallback rules would work in practice, or how quickly the system could adapt.

The central lesson from this learning process is that reducing strategic ranking incentives is necessary but not sufficient. Even under a low-strategy assignment mechanism, families still need to discover schools, compare quality and cost, understand admission chances, and realize when their current application portfolio exposes them to non-placement risk. The evidence reviewed in the paper shows that families often overestimate admission chances, underestimate feasible nearby options, and stop searching too early. These frictions are not merely informational details; they affect who can use the assignment system effectively.

SAE’s response was to develop a “smart platform” layer on top of the assignment mechanism. Risk warnings, personalized feedback reports, the School Explorer/MIME tools, WhatsApp nudges, and later Anótate en la Lista all use platform data to make options, risks, and consequences more visible to families. The evidence points to a consistent pattern: platform-embedded information can change application behavior. Families add options, reduce predicted non-placement risk, and in several studies move toward schools with higher measured value-added. New follow-up evidence is consistent with some of these gains translating into later learning outcomes, though the paper treats those results conservatively and distinguishes published evidence from new and companion working-paper results.

The paper also adds three empirical contributions to the existing evidence. First, it reports an out-of-sample replication of the risk-warning regression discontinuity after the original QJE study window, covering SAE cycles from 2018 to 2023 and linking later cohorts to fourth-grade SIMCE outcomes. Second, it uses linked administrative records to quantify the new in-year aftermarket platform, Anótate en la Lista, showing that off-cycle applications are comparable in scale to the main round but still operate largely outside SAE’s equity priorities. Third, it presents staggered event-study evidence on segregation, showing convergence in prioritario representation at oversubscribed schools, where the assignment mechanism actually binds.

After a decade, SAE’s aggregate assignment performance is strong but bounded. In 2025, 82.6 percent of applicants were assigned to one of their top three listed schools and 92.8 percent to a listed preference. These outcomes place Chile within the range of mature centralized systems. But the remaining 7.2 percent non-placement rate is geographically concentrated, and national averages hide local markets where desired seats are scarce. The evidence is consistent with local supply constraints shaping what SAE can deliver: demographic growth and migration increased demand in some areas, private-subsidized school entry slowed sharply, and some desirable high-value-added private-subsidized schools exited to the unsubsidized private sector. As a result, empty seats in some markets coexist with acute shortages in others.

The equity effects are similarly bounded. SAE appears to have improved access at oversubscribed schools, where discretionary screening previously mattered most and where the allocation rule has real bite. However, aggregate segregation effects are harder to identify causally and depend on the measure. Residential stratification, the unsubsidized private sector, school entry and exit, and local supply constraints all operate outside the assignment algorithm. The right interpretation is therefore neither triumphalist nor dismissive: SAE delivered measurable gains in access at constrained, high-demand schools, while leaving broader segregation and quality-access problems for complementary policy.

The unfinished agenda follows directly from these limits. First, assignment policy must be connected more explicitly to supply policy. A centralized mechanism can allocate scarce seats more transparently, but it cannot create desirable seats where too few exist. Second, the aftermarket needs stronger design. Anótate en la Lista made in-year demand visible at scale, but its first-come-first-served logic does not yet incorporate the equity priorities of the main round. Third, many algorithmic and priority-rule adjustments remain to be debated and evaluated, including sibling logic, fallback rules, tie-breaking, capacity provisions, and priority definitions. Fourth, communication with families remains a policy margin: families need clearer, timely, personalized information about risk, feasible options, sibling trade-offs, fallback consequences, and aftermarket rules.

The broader message is that Chile’s experience is best understood as institution plus learning. SAE built a transparent national assignment system, used staggered implementation to learn while scaling, and developed smart-platform tools to address the behavioral frictions revealed by that process. The next decade requires extending the same iterative discipline to the margins that have changed more slowly: supply, the aftermarket, algorithm and priority adjustments, communication with families, and public governance of complex assignment rules. Centralized assignment can make school access fairer, more transparent, and more empirically governable, but the mechanism is only one part of the institution. Its long-run value depends on whether the system can keep learning from its own operation and translate that learning into rules, tools, and policies that families can trust.