SG-D-36 | Education Streaming Reform: From Streaming to Subject-Based Banding (1980–2026)

---

---

1. Key Takeaways

• The 1979 Goh Report was the founding act of Singapore's streaming era. Commissioned by Deputy Prime Minister Goh Keng Swee, the report documented catastrophic educational wastage — fewer than 60 percent of each primary school cohort could meet minimum literacy standards under the then-undifferentiated system. Its remedy was unambiguous: sort students by ability at the end of Primary Three into differentiated language streams (EM1, EM2, and EM3), each with different curricular pace and examination demands. This act of administrative sorting, operationalised from 1981, transformed Singapore's aggregate educational outcomes within a decade. It also created the social architecture of early labelling that would define the life trajectories of hundreds of thousands of Singaporeans for four decades.

• The EM3 stream became the most contested institution in Singapore's social policy history. Officially a "slower pace" track designed to ensure that academically weaker students received instruction tailored to their current level, EM3 in practice functioned as an early dead-end. Students placed in EM3 — disproportionately from lower-income families and ethnic minority backgrounds — were streamed at age nine into a track from which exit was difficult and whose PSLE ceiling foreclosed access to the Express secondary stream. The EM3 label carried intense stigma; educators, parents, and later critics including Teo You Yenn documented how EM3 children internalised the categorisation as a statement about their fundamental worth. The Ministry of Education abolished EM3 in 2003 under Education Minister Tharman Shanmugaratnam — the first structural acknowledgement that the 1979 sorting architecture had costs the system could no longer absorb.

• The 2004–2021 period was a sustained, incremental dismantling of the T-score regime. The PSLE T-score — a standardised aggregate score that ranked each cohort on a bell curve — was the mechanical heart of secondary school placement. Its aggregate nature meant that a student's PSLE T-score could be raised or lowered by performance in a single subject, producing extreme sensitivity to narrow differences and feeding the anxiety-driven tuition industry. The shift from T-scores to Achievement Levels (AL), announced in 2016 and implemented from the 2021 Primary Six cohort onwards, replaced continuous ranking with eight broad bands per subject, aggregated into a total AL score. This change reduced extreme differentiation without eliminating examination-based placement.

• Full Subject-Based Banding (FSBB), implemented from 2024, represents the most fundamental restructuring of secondary education since the streaming system itself. Under FSBB, the Express, Normal Academic, and Normal Technical stream labels were abolished. Instead, students are offered each subject at one of three levels — G1 (General 1, broadly equivalent to the former Normal Technical), G2 (General 2, broadly equivalent to Normal Academic), and G3 (General 3, broadly equivalent to Express). A student can take different subjects at different G-levels — mathematics at G3 while taking English at G2, for instance — and progress to higher G-levels within a subject over time. The system is designed to eliminate stream-based identity and the fixed ceiling that stream placement previously imposed.

• The Gifted Education Programme (GEP) and Integrated Programme (IP) have been deliberately insulated from streaming reform. While the mainstream structure has been progressively liberalised, the elite tracks — GEP (selection at Primary Three), IP (allowing top secondary schools to offer a six-year combined O-Level/A-Level equivalent programme), and Special Assistance Plan (SAP) schools — have been preserved and modestly restructured rather than dismantled. The 2024 GEP restructure moved enrichment activities to a broader set of schools rather than concentrating them in three dedicated schools, but maintained the selection mechanism. These elite pathways remain the most durable features of the stratified system.

• The critiques from Singapore's critical social sciences have been systematic and largely unrefuted on the data. Teo You Yenn's This Is What Inequality Looks Like (2018) documented how the streaming system concentrated disadvantage and made it illegible as structural. Donald Low and Sudhir Thomas Vadaketh's Hard Choices (2014) situated streaming within a broader analysis of Singapore's meritocratic consensus as ideological cover for class reproduction. Kenneth Paul Tan's academic work traced how the ideology of meritocracy evolved to naturalise outcomes that were substantially shaped by prior advantage. None of these critiques have been directly refuted by the government; instead, each wave of reform has partially incorporated their empirical observations while preserving the competitive architecture on which the selection system depends.

• The international comparisons are instructive but contested. Singapore consistently ranks at or near the top of PISA and TIMSS international assessments. Proponents cite these rankings as evidence that the system, including streaming's successors, is effective. Critics note that PISA scores measure tested knowledge under examination conditions and do not capture wellbeing, creativity, anxiety levels, or the distributional consequences of how average high performance is achieved. Singapore's PISA scores partly reflect a system that selects and concentrates test-optimised students into a small number of elite schools — a form of average elevation that does not contradict the existence of sharp inequality at the tail.

• By 2026, the streaming architecture has been substantially dismantled in formal terms but its social consequences persist. The AL scoring system, FSBB, and the restructured GEP represent genuine structural changes. But the private tuition industry, the concentration of resources in historically elite schools, the SAP school ethnic composition question, and the continued operation of IP schools within the same system mean that the underlying competitive sorting pressure has not dissipated — it has been rerouted. The central tension between meritocratic selection and equitable opportunity remains unresolved and is likely to define education policy debates well into the 2030s.

---

2. The Record in Brief

Singapore's management of educational streaming is the story of a system that was built to solve one problem, succeeded beyond its designers' expectations, created a new and more insidious problem, and has been methodically dismantled — section by section — over four decades, without ever being fully replaced. The arc from Goh Keng Swee's 1979 Report on the Ministry of Education to the 2024 full implementation of Subject-Based Banding encompasses some of the most consequential social engineering decisions in Singapore's domestic governance, decisions made under the legitimating banner of meritocracy but with consequences that have been documented, contested, and slowly acknowledged as inequitable by the very government that designed them.

The colonial inheritance was an education system that was neither universal nor coherent. British Malaya had treated education primarily as a matter for communal management — the Chinese community built and funded Chinese-medium schools, the Muslim community operated Malay-medium schools, and the small English-medium sector trained a colonial administrative class. Tamil-medium education served plantation workers. These streams converged on nothing: their graduates could not communicate across linguistic lines, their examination systems were incomparable, and their political loyalties diverged accordingly. The Chinese-medium sector, in particular, was a major site of left-wing organisation; the Nanyang University student body was one of the PAP's most difficult political constituencies in the 1950s.

The PAP's first decade and a half was consumed by the bilingual policy project — establishing English as the common medium of instruction while preserving mother-tongue languages as compulsory second languages. This was a project documented in full in SG-A-16 and need not be rehearsed here. By the late 1970s, the medium-of-instruction question had been substantially resolved: English-medium enrolment had overtaken Chinese-medium, and the separate-stream system was dying. What the government then confronted was a unified system that was failing on aggregate measures.

The 1979 Goh Report provided the statistical shock. Under the unified system, approximately 40 percent of each primary school cohort was leaving primary school without adequate literacy in either English or their mother tongue — the "10-year wastage" that Goh Keng Swee identified as the system's central failure. Students were being held back in an undifferentiated stream that moved too fast for weaker learners and too slowly for stronger ones. The prescribed remedy was sorting: stream students into differentiated ability groups, with each group receiving instruction at an appropriate pace and level. The policy logic was economistic and uncompromising — maximise the productivity of each cohort by eliminating the mismatch between instructional pace and student capacity.

The streaming architecture that resulted — EM1, EM2, and EM3 at primary level, Express and Normal streams at secondary — was operationalised from 1981. Its aggregate outcomes were, by its designers' own metrics, a success. Cohort completion rates rose dramatically; the proportion of students achieving minimum literacy standards improved substantially within a decade; and Singapore's rankings in international educational assessments began their ascent. By 1990, Singapore was a recognised case study of rapid educational improvement.

But the system's social architecture was producing consequences that were visible to anyone who looked. Stream placement correlated with socioeconomic status. Children of professional households were disproportionately placed in EM1; children of lower-income households were disproportionately placed in EM3. Stream placement at age nine functioned as a self-fulfilling prediction: EM3 students were taught a narrower curriculum, received less intensive academic preparation, and emerged from primary school with fewer options at the secondary level. The secondary Normal Technical stream, the terminal consequence of EM3 placement, was colloquially known as a "dead end." The children in it knew it. So did their teachers.

The government was not unaware of these criticisms. Minister for Education Goh Chok Tong acknowledged in parliamentary debates as early as the mid-1980s that the system produced stigma. But the structural response was slow. The political economy of reform was difficult: streaming was popular with parents of higher-streamed children, who understood that their children were receiving better-resourced instruction. Any dilution of streaming was experienced by this politically active constituency as a threat. The system persisted with incremental modifications — additional pathways, bridging mechanisms, the introduction of a Normal Academic versus Normal Technical distinction at secondary level in 1991 — but without structural dismantlement.

The 1997 Goh Chok Tong "Thinking Schools, Learning Nation" speech signalled the beginning of a sustained ideological shift. What had been presented as a matter of scientific sorting began to be presented as a matter of values: Singapore wanted not only efficient processors of knowledge but creative, curious, intrinsically motivated learners. The tension between these two goals — a sorting machine and a creative academy — has never been fully resolved; the system has attempted to do both, with mixed results. But from 1997 onward, the direction of reform was consistently toward greater flexibility, reduced labelling, and broader banding.

The policy chronicle from 2003 to 2026 is one of progressive structural reform, each step announced as a response to diagnosed problems, each step opposed by the organised interests of parents invested in the existing system, and each step taken with careful attention to the political optics of meritocracy. The government has never conceded that streaming was a mistake. The official framing, consistent through Tharman, Heng Swee Keat, and Chan Chun Sing as successive Education Ministers, has been that streaming served its purpose in an earlier era, that Singapore's students and teachers have evolved beyond the system, and that the reforms represent not an admission of error but an upgrading of a system that had accomplished its original objectives.

This framing is contestable. The research evidence, reviewed in Sections 9 and 11, suggests that the system's design choices — particularly the early-age sorting, the T-score aggregate, and the stream-based identity — created path dependencies that were costly and that the government declined to acknowledge as policy failures. What is clear is that by the mid-2010s, the costs of the streaming architecture had become politically salient in ways they had not been in the 1980s and 1990s, and that the reforms of 2019–2024 represent the most significant structural response to those costs since the system was built.

---

3. Timeline 1980–2026

---

4. The 1979 Streaming Report (Goh Keng Swee) and the Original EM1/EM2/EM3 Architecture

The Report on the Ministry of Education 1978, submitted by Goh Keng Swee and the Education Study Team in late 1978 and published in 1979, is the foundational document of Singapore's streaming era. To understand why the streaming system took the particular form it did — and why its dismantlement has been so incremental — it is necessary to read the Goh Report not merely as a pedagogical document but as a political-economic intervention by a minister who approached education primarily through the lens of human capital productivity.

Goh Keng Swee's diagnosis was statistical and unsparing. Under the system as it existed, approximately 39 percent of students who began primary school did not complete ten years of formal education; of those who did complete primary school, a substantial proportion emerged without adequate English and mother-tongue literacy. Goh framed this as "wastage" — a term that was deliberately economic in its connotation, treating uncultivated educational potential as a cost to the national enterprise. The language was deliberate: education reform in Singapore has consistently been presented as economic necessity rather than social equity, and the Goh Report established this framing definitively.

The report's central recommendation was the differentiation of instruction into three streams based on assessed language ability at the end of Primary Three (approximately age nine). Students would be placed in EM1 (English-medium instruction in both English and a second language, with the most demanding pace), EM2 (English-medium instruction with a slower pace in a second language), or EM3 (instruction primarily in the vernacular, with English as a second language, at the slowest pace). Placement would be based on continuous assessment during Primary One to Three. Movement between streams was possible in principle, though the mechanisms for upward movement were limited in practice.

The rationale was pedagogically coherent given the assumptions. If students learn at different rates, a single instructional pace will be too fast for some and too slow for others; differentiated instruction allows teachers to focus on a narrower ability range and progress at an appropriate pace. The system was borrowed, loosely, from British grammar-school-era thinking about ability grouping, though Singapore's version was more rigid in its early years and more consequential in its downstream effects.

The original architecture functioned as follows. At the end of Primary Three, students were assessed and placed in a stream. EM1 students studied two languages intensively and had access to the full secondary curriculum. EM2 students studied two languages at a slower pace but could still access the secondary Express stream if their PSLE performance was strong. EM3 students were taught primarily in their mother tongue, with English as a second language — a design that reflected the government's concern that some students were being overwhelmed by the demands of the bilingual policy — and whose PSLE ceiling made access to the Express secondary stream very unlikely. ⚠Requires verification.

The immediate academic consequences were, by the government's measures, positive. Repetition rates declined. Completion rates rose. The proportion of students achieving basic literacy targets improved. By the late 1980s, Singapore's primary school system had moved from near-crisis levels of wastage to respectable aggregate performance. These improvements were real and are not seriously disputed by critics of the streaming system. The debate is not about whether streaming improved average outcomes but about the distributional costs and the social consequences of early sorting.

Those consequences emerged clearly in the research record of the 1980s and 1990s, though they were acknowledged politically only slowly. Several patterns were documented. First, stream placement correlated with socioeconomic status: children of professional and higher-income families were more likely to be in EM1; children of lower-income families, particularly those in public rental housing with parents who had limited formal education, were more likely to be in EM3. This correlation was not surprising — early-age language assessments inevitably reflect the linguistic environment at home, and richer homes provide richer linguistic environments — but it meant that the streaming system was not a neutral sorting of innate ability but a sorting of prior advantage, legitimated by the language of ability.

Second, stream placement functioned as a self-fulfilling categorisation. Students in EM3 were taught a narrower curriculum by teachers who had, in many schools, lower expectations of their performance. The ceiling implicit in the stream label was absorbed by teachers, parents, and students themselves. Ethnographic research — limited by Singapore's tight constraints on independent social research in the 1980s and 1990s — pointed to the psychological costs of EM3 placement, particularly for Chinese students whose parents had themselves been educated in the Chinese-medium stream and whose EM3 placement felt like a double degradation: not only academically categorised as the weakest, but taught through the medium their parents' generation had been told was inferior.

Third, ethnic composition varied across streams in ways that mapped onto existing socioeconomic patterns. Malay students were disproportionately placed in EM3, reflecting the socioeconomic gaps between communities that were themselves partly the legacy of colonial-era occupational segregation. Tamil-medium instruction had effectively collapsed by the late 1970s; Indian students were being absorbed into the English-medium system at different rates. ⚠Requires verification.

Goh Keng Swee was aware, from correspondence and parliamentary exchanges, that the system would produce these distributional consequences. His response, consistent with his overall governing philosophy, was essentially utilitarian: the system maximised aggregate educational productivity; the distributive consequences were a cost worth paying; and targeted remedial mechanisms could address the worst cases. This position held for two decades, through successive Education Ministers who tinkered with the margins of the system without questioning its architecture.

The secondary level streaming architecture established in 1983 mirrored the primary structure. Students exiting primary school were placed in one of two secondary streams — Express (leading to O-Levels in four years) and Normal (five years, leading to N-Level examinations, with O-Level access only for top Normal students). The Normal stream was subsequently bifurcated into Normal Academic and Normal Technical in 1991, creating a three-way secondary structure that would persist until 2024. The secondary streaming system had consequences analogous to the primary: Normal Technical students were channelled toward Institute of Technical Education vocational training, while Express students accessed junior colleges and universities. The differential secondary pathways were not merely educational in consequence; they were life-trajectory defining in a credential-intensive society.

---

5. The 2003–2008 Reforms: Abolition of EM3 and the Tharman Era

Tharman Shanmugaratnam's appointment as Education Minister in 2003 marked the beginning of the most sustained reformist period in Singapore education policy since the Goh Report itself. Tharman brought an intellectual framework that differed from his predecessors: trained in economics at LSE and Cambridge, and later at the IMF, he approached education not as a human capital sorting machine but as the foundational mechanism of social mobility. His central conviction — articulated in Committee of Supply speeches, public lectures, and later in international forums as Finance Minister and Deputy Prime Minister — was that meritocracy, if left unreformed, calcifies into a system of inherited privilege. The streaming system, in his analysis, was the most visible mechanism of this calcification.

The first structural decision was the abolition of EM3 for new Primary One cohorts from 2003. Students who had entered EM3 in earlier years completed their primary education in the existing system, but no new placements were made. From 2008, EM3 had been entirely phased out of the primary school system. The official framing was that improvements in teaching capacity, bilingual policy adjustments, and changes to the mother-tongue curriculum had made the EM3 differentiation unnecessary — weaker learners could be accommodated through differentiated pace within a unified stream. The implicit acknowledgement was that EM3's costs in stigma and ceiling-setting had become unacceptable.

Simultaneously, Tharman introduced the Subject-Based Banding pilot at the primary level in 2004. Under this pilot, a limited number of primary schools offered certain subjects — initially English and mathematics — at different levels, allowing students to take, say, Standard English while taking Higher Mathematics. The pilot was limited in scope and received relatively little public attention at launch, but it established the conceptual precedent that would become the full FSBB architecture fifteen years later: that a student need not be assigned to a single stream but could take different subjects at different levels of difficulty.

The "Teach Less, Learn More" initiative, formalised in 2004, pushed in a complementary direction: reducing the emphasis on examination drilling and increasing time for inquiry-based learning, discussion, and student-directed activity. Whether TLLM changed classroom practice in a durable way is contested — surveys of teachers in the late 2000s suggested that the examination imperative continued to dominate instructional time — but as a statement of values, it positioned the government against pure credential-racing, an important signal in a system where the Tharman critique of gaming-driven learning was well-founded.

Heng Swee Keat, who succeeded Tharman as Education Minister in 2011, continued the reform trajectory with a focus on the PSLE itself. The PSLE's T-score aggregate had been the mechanical driver of the most dysfunctional aspects of the system. Because the T-score was a standardised score relative to the cohort — not an absolute score — a student's result depended on how well the entire cohort performed. This created an incentive structure that was collectively irrational: any individual family that reduced tuition time while others maintained it would fall behind in relative terms, even if absolute knowledge had improved. The tuition industry's extraordinary scale — estimated by the 2010s at over S$1.4 billion annually, with participation rates of 70–80 percent at primary school level — was partly a rational response to this competitive dynamic.

The T-score also produced extreme sensitivity to narrow differences. A difference of one PSLE aggregate point could determine whether a student gained entry to a specific secondary school. This sensitivity drove extreme preparation, high parental anxiety, and — crucially — the distortion of secondary school choice toward obsessive score-target calculation rather than genuine preference for school culture or curriculum.

Heng announced in 2012 that MOE was conducting a comprehensive review of the PSLE. The process was unusually participatory by Singapore standards: public forums, parent surveys, and structured consultations were held over three years. The outcome, announced in 2016, was the replacement of T-scores with Achievement Levels (AL) — eight broad bands per subject (AL1 being the highest, AL8 the lowest), with the PSLE aggregate being the sum of four subject AL scores. Implementation was set for the Primary Six 2021 cohort, giving schools and families five years to adjust.

---

6. The 2019 Subject-Based Banding Decision and Full Implementation 2024

The announcement in November 2019 by Education Minister Ong Ye Kung of Full Subject-Based Banding (FSBB) represented the structural culmination of a reform trajectory that had been building since Tharman's abolition of EM3. Where previous reforms had modified the assessment system or introduced marginal flexibility at the edges, FSBB abolished the stream-based secondary architecture entirely.

The architecture of FSBB, as announced and subsequently refined, replaced the three secondary streams — Express, Normal Academic, and Normal Technical — with three subject-level designations: G1, G2, and G3. These designations correspond, in terms of content and pace, roughly to the former Normal Technical, Normal Academic, and Express levels respectively. The critical departure from the old system is that the designations are subject-specific rather than stream-wide. A student is not placed in "G2" as a global category; rather, a student might take English at G3, Mathematics at G2, and a science subject at G1, depending on their PSLE performance in each subject and subsequent school-year progression.

The official rationale was articulated by MOE in terms of recognising diverse strengths, reducing the stigma of stream labels, and enabling more permeable pathways between subjects and levels. A student who entered Secondary One with G1 Mathematics could, over time, demonstrate sufficient mastery to progress to G2 or G3 in that subject. The system was designed to make progression possible without requiring a wholesale stream transfer — the mechanism that had theoretically allowed movement between streams in the old system but which was rarely used and carried high costs in terms of curricular disruption.

The phased implementation timeline was as follows. Secondary schools began piloting FSBB elements from 2022, with mixed-level classes in certain subjects. From 2023, FSBB was expanded to all mainstream secondary schools. Full implementation — with stream labels removed and G1/G2/G3 universally applied — was achieved for the 2024 Secondary One cohort. Students who had entered the system under the previous stream architecture completed their secondary education under transitional arrangements; the final cohort under the old streaming system completed their N-Level or O-Level examinations by 2025.

The transition presented implementation challenges that MOE addressed through detailed guidance to schools and teachers. Mixed-level subject classes required teachers to manage classrooms with students from different G-levels working on different content — a significant pedagogical demand that required substantial teacher development investment. Whether the teacher workforce was adequately prepared for this shift, and whether the classroom practice reality matched the policy design, are questions that the Ministry's own evaluation mechanisms were tracking through the 2023–2025 period. ⚠Requires verification.

The secondary school landscape under FSBB remained stratified in practice, though in different ways than before. Secondary school admission continued to be based on PSLE AL scores, meaning that schools maintained implicit score profiles associated with their student bodies. A school historically associated with the Express track would continue to attract students with lower (better) PSLE AL aggregates, while a school historically associated with Normal Technical would attract students with higher aggregates. The G-level composition of a school's student population would, under FSBB, reflect these prior differentials. The label had changed; the underlying sorting had not.

The SAP (Special Assistance Plan) schools — nine secondary schools that offer both English and Higher Chinese and maintain Chinese cultural traditions — were not restructured under FSBB. They continue to admit students based on PSLE aggregate and Higher Chinese performance, and their student bodies remain Chinese-medium by design. The ethnic composition question, discussed in Section 10, remained the most politically sensitive unresolved issue in the FSBB transition.

---

7. The 2021–2024 PSLE Achievement Level Scoring System

The replacement of T-scores with Achievement Levels constitutes the most consequential assessment reform in Singapore's primary education history since the PSLE itself was established as the universal primary school leaving examination. Understanding the mechanics of the AL system is essential to evaluating whether the reform achieved its stated objectives.

Under the former T-score system, each subject score was standardised relative to the cohort performance in that subject — producing a score that measured not absolute knowledge but relative standing within the year group. A student who scored 85 out of 100 in mathematics might receive a T-score of 60 or 55 depending on whether the cohort as a whole had scored higher or lower that year. The four subject T-scores were aggregated into a total, and secondary school admission was based on this aggregate, with fine differentiation producing intense competition for small score differences.

Under the AL system, each subject is scored into one of eight bands: AL1 (highest, roughly corresponding to a very high raw score), AL2, AL3, AL4, AL5, AL6, AL7, and AL8 (lowest). ⚠Requires verification. The PSLE aggregate is the sum of four subject AL scores, ranging from a minimum of 4 (four AL1s) to a maximum of 32 (four AL8s). Lower aggregates are better. A student with four AL1s has the best possible aggregate of 4; a student with four AL8s has 32.

Secondary school admission is based on the PSLE aggregate, with schools having score ranges (lower is more competitive). Because AL bands are broad — a range of raw scores maps to the same AL — two students with raw scores of 87 and 81 in a subject might both receive AL2, whereas under T-scores, the six-point difference would have produced a meaningful T-score differential. This was the reform's design objective: reduce fine differentiation, reduce the marginal value of additional tuition aimed at moving from, say, a raw score of 82 to 85, and correspondingly reduce examination anxiety.

The evidence from the first cohorts to go through the AL system — the 2021 Primary Six cohort who entered secondary school in 2022, and subsequent cohorts — suggested that the reform had partially achieved its objectives. MOE surveys indicated that parental anxiety about PSLE decreased somewhat in the AL era compared to the T-score era. The tuition industry, however, showed no evidence of contraction in the 2022–2025 period; aggregate sector revenue continued to grow, suggesting that while the AL reform may have reduced the marginal return to fine-grained drilling, the overall competitive pressure that drives tuition demand persisted. ⚠Requires verification.

The AL system was also designed to reduce the importance of the PSLE relative to the student's secondary school journey. Under the new secondary framework (FSBB), students were to be placed not by a single PSLE aggregate into a fixed stream but by subject-level performance into G-levels that could subsequently change. The PSLE AL aggregate would determine initial secondary school admission and initial G-level placement, but it would not fix a ceiling. A student with a G2 placement in mathematics at Secondary One who demonstrated strong performance could, in principle, move to G3. This permeability — limited in the old system to formal stream transfer requests that were rarely approved — was intended as the structural complement to the AL reform.

Whether the AL system, combined with FSBB, has produced genuinely more permeable pathways is a question the data through 2026 is only beginning to answer. The 2024 Secondary One cohort — the first to enter under full FSBB — will not produce meaningful outcome data until they complete their secondary education and proceed to post-secondary institutions, which will not be observable until 2028 at the earliest. The policy design is structurally sound; its empirical effectiveness remains to be verified.

One dimension of the AL reform that critics found unsatisfying was the continued existence of the secondary school admission ranking. Even under AL, different secondary schools had different score profiles, and the schools historically associated with the Express track continued to attract the lowest-aggregate (most competitive) students. This maintained the underlying school hierarchy — which has significant consequences for teachers, resources, networking, and post-secondary pathways — while removing the T-score apparatus that had most visibly and mechanically operationalised that hierarchy. Whether abolishing the label while preserving the substance constitutes genuine reform or cosmetic adjustment is one of the most hotly contested questions in Singapore's education policy discourse as of 2026.

---

8. The Special/Express/Normal Reform — Phased Replacement with FSBB

The phased replacement of the Express/Normal Academic/Normal Technical stream labels with FSBB's G1/G2/G3 subject-level designations was the most operationally complex component of the reform programme. It required simultaneous changes to teacher training, curriculum materials, timetabling, classroom organisation, parental communication, and the public discourse around educational expectations. The challenge was not merely administrative; it was cultural. The stream labels — Express, Normal Academic, Normal Technical — had become deeply embedded in social language, carrying associations that went well beyond their technical definitions.

"Express" connoted academic success, university eligibility, and access to the professional class. "Normal Technical" connoted academic struggle, manual trades, and — in the bluntest lay readings — limited future prospects. These associations were not imaginary projections; they reflected real statistical patterns in where stream graduates ended up in the labour market and in life. Students in Normal Technical were disproportionately channelled toward ITE (Institute of Technical Education) and lower-wage technical occupations. Students in Express were disproportionately channelled toward junior colleges, universities, and professional careers. The label was not merely descriptive; it was a compound of description and prediction.

The government's decision to remove the labels was explicitly motivated by the desire to break the psychological foreclosure that the labels created. Education Ministers from Tharman onward had repeatedly cited evidence that students in Normal Technical streams internalised the label as a statement about their intelligence and potential, and that this internalisation — what social psychologists would describe as stereotype threat — suppressed performance and aspiration. The label removal was an attempt to disrupt this feedback loop.

The mechanics of the transition were carefully sequenced. From 2022, mixed-level subject groupings began appearing in schools that had volunteered for FSBB pilots. From 2023, all mainstream secondary schools began implementing mixed-level groupings in at least some subjects. The 2024 Secondary One cohort entering in January 2024 was the first cohort to experience a fully label-free secondary school environment from the outset. For students already in the system under the old stream labels, their course completion was managed under transitional arrangements that preserved their existing course choices while aligning their subsequent subject selections with the G-level framework where feasible.

The practical challenge of mixed-level classrooms deserves detailed attention. Under the old streaming system, a secondary school English class contained students of approximately similar assessed English ability, which simplified lesson planning. Under FSBB, a class might contain students registered for G2 and G3 English — or even G1 and G2 in other subjects — creating a wider ability spread that requires differentiated instruction within the classroom. This is a common approach in many education systems, but Singapore's teacher workforce had been trained and developed within a system that relied on relatively homogeneous stream-based classes. The professional development investment required to equip teachers for differentiated instruction — and the time costs of delivering such instruction — were real and not negligible. ⚠Requires verification.

The resource allocation implications were also managed deliberately. Historically, Singapore's streaming system had created resource stratification: schools that attracted Express-stream students had higher parental socioeconomic status, stronger alumni networks, greater capacity for fundraising, and higher demand from experienced teachers. FSBB's removal of stream labels could not, by itself, address this structural resource differential; it required complementary policies to ensure that schools historically associated with Normal streams received adequate support. MOE's stated commitment to resourcing schools on an equal basis was articulated through multiple ministerial statements, but the structural advantages of elite schools — teacher quality, alumni networks, co-curricular resources — are difficult to equalize through resource allocation alone.

The 2024 implementation also intersected with the broader restructuring of post-secondary pathways. Students completing their secondary education under FSBB would take either the Singapore-Cambridge Secondary Education Certificate (equivalent to the former O-Level for G3 subjects) or mixed qualifications depending on their G-level choices. The post-secondary system — junior colleges, polytechnics, ITE, applied learning programmes — was simultaneously being adapted to admit students on the basis of their G-level subject performance profiles rather than stream-based O-Level aggregates. This represented a substantial administrative restructuring of the post-secondary admission framework that was still being finalised in the 2024–2025 period. ⚠Requires verification.

---

9. The Critiques — Teo You Yenn, Donald Low, Kenneth Paul Tan

The academic and public intellectual critique of Singapore's streaming system has been substantial, sustained, and — in the estimation of many observers — empirically persuasive even where politically marginalised. Three bodies of work stand out as defining contributions to this literature.

Teo You Yenn is a sociologist at Nanyang Technological University whose book This Is What Inequality Looks Like (2018, Ethos Books) became the most widely read work of critical social analysis published in Singapore in recent memory. While not exclusively about education streaming, Teo's work is saturated with its consequences. Her ethnographic research among low-income families in rental housing documented how the streaming system operated as a mechanism of social invisibility: children placed in Normal Technical were simultaneously categorised as academically weak, channelled toward lower-wage occupations, and rendered less legible to the policymakers and upper-income Singaporeans who design and inhabit the dominant institutions of Singapore life. Teo's central argument about inequality is that Singapore's ideological framing of social outcomes as the products of individual effort and merit makes structural disadvantage personally invisible — the streaming system is a prime example of a structural sorting mechanism presented in the language of individual ability assessment.

Teo has been careful not to argue against all ability grouping in principle; her critique is specifically about early sorting, rigid labels, and the ideological work done by presenting socially stratified outcomes as the products of meritocratic measurement. Her work resonated with a broad educated readership in Singapore in ways that earlier academic critiques had not, partly because of its accessible prose and partly because it was published at a moment when income inequality and housing affordability had become salient political anxieties. The book's sales figures — remarkable for academic social analysis — indicate that its diagnosis found a receptive audience.

Donald Low and Sudhir Thomas Vadaketh's Hard Choices: Challenging the Singapore Consensus (2014, NUS Press) approached streaming through the lens of political economy. Low, a former senior civil servant at the Ministry of Finance and later a professor at HKUST, argued that streaming was one component of a broader ideological consensus — the "Singapore consensus" — that used the language of meritocracy and pragmatism to insulate a set of policy choices from critical examination. His critique of streaming was not merely that it produced inequitable outcomes but that the combination of early sorting, high-stakes examination, and meritocratic legitimation prevented honest public debate about the system's design choices. Low's argument was that Singapore's technocratic governance model was comfortable commissioning evidence about educational outcomes but resistant to the political consequences of acting on evidence that challenged the fundamental streaming architecture.

Low also drew attention to the distinction between equality of opportunity (which the streaming system claimed to provide through examination-based sorting) and genuine equalisation of starting positions (which the streaming system manifestly did not provide, given the correlations between socioeconomic background and pre-school linguistic environment). This distinction — common in philosophical analyses of distributive justice — had been largely absent from Singapore's official discourse about meritocracy and education, and Low's articulation of it helped frame subsequent debates.

Kenneth Paul Tan, a political scientist who has held positions at NUS Lee Kuan Yew School of Public Policy and Yale-NUS College, has written extensively on meritocracy as ideology. His 2008 article in International Political Science Review argued that Singapore's concept of meritocracy had undergone an ideological shift from its original egalitarian promise — the idea that talent from any background could rise — toward a more elitist formulation that naturalised the concentration of privilege in high-performing schools and scholarship pipelines. Tan's reading of streaming was as the lower end of this elitist meritocracy: the streaming system not only selected the best but — by labelling and routing the rest — actively reproduced the class structure it claimed to sort by merit alone.

Tan's subsequent work has engaged with Forward Singapore and the reforms of the 2019–2024 period, arguing that the removal of stream labels constitutes a partial but insufficient response to the structural critique. His position, as of 2025, is that FSBB addresses the most visible and most symbolic aspects of the streaming problem — the labels, the rigidity, the early sorting at age nine — without addressing the underlying mechanisms: the differential school quality, the tuition industry amplifier, and the continued operation of the GEP and IP elite pathways within the same system.

The government's response to these critiques has been nuanced. Ministers have consistently acknowledged the equity concerns raised by these scholars while disputing the more radical policy prescriptions. The official line — articulated most articulately by Tharman in his Amartya Sen Lecture (2015) and by Chan Chun Sing in FSBB-related parliamentary speeches — is that meritocracy is not the problem but its incomplete implementation: if the meritocracy is not working for some children because they lack equal preparation, the response is to improve their preparation (through subsidised preschool, financial support for lower-income families, direct school support schemes) rather than to abandon the examination-based sorting that the system requires to identify where to invest talent.

This is a coherent position. Whether it is adequate — whether investment in early childhood and remedial support can offset the structural advantages that wealthier families compound through private tuition, cultural capital, and social networks — is the empirical question that will define Singapore's education policy debate in the coming decade.

---

10. The IB / IP / Gifted Education Programme Tracks

The Integrated Programme (IP) and Gifted Education Programme (GEP) represent the elite tier of Singapore's secondary education system, and their relationship to the streaming reform is one of studied preservation rather than dismantlement.

The Gifted Education Programme was launched in 1984 on the recommendation of the then-Education Ministry, influenced by international research on the underserved needs of highly able students and — less directly acknowledged — by Lee Kuan Yew's longstanding conviction that Singapore's most precious resource was its small cohort of exceptional cognitive talent. The GEP selects students at the end of Primary Three through a two-stage screening and assessment process, identifying approximately ⚠Requires verification of each cohort. GEP students receive an enriched curriculum in three designated primary schools: Nan Hua Primary, Raffles Girls' Primary, and Anglo-Chinese Primary.

The 2024 GEP restructure, announced by MOE and implemented progressively from that year, modified the delivery model: GEP enrichment activities and curriculum elements are being distributed to a broader set of schools beyond the three dedicated GEP schools. The selection mechanism — the Primary Three screening — is retained, as is the designation of students as GEP-selected. The rationale for the change was to reduce the geographical and social concentration of GEP students in three specific schools, which had created a feeder pipeline to specific secondary schools and reinforced existing patterns of elite school selection. Whether distributing GEP activities more broadly, while retaining the selection label, meaningfully changes the social dynamics of elite formation is contested. Critics note that the GEP identity — and the school choice decisions made around it — may be more durable than the geographical distribution of the programme.

The Integrated Programme (IP), introduced in 2004, allows students in the highest-performing secondary schools to skip the O-Level examination and proceed on a six-year course to the A-Levels (or, in some schools, to the International Baccalaureate). IP schools include Raffles Institution, Hwa Chong Institution, Anglo-Chinese School (Independent), Nanyang Girls' High School, National Junior College (combined with its feeder secondary), and several others. The IP pathway was designed to free these schools from the constraints of O-Level examination preparation and allow more creative, project-based learning in the secondary years.

The IP has achieved some of its educational objectives — IP schools consistently produce A-Level and IB results that are strong by international standards — but it has also reinforced the concentration of talent and resources in a small number of schools. Admission to IP schools is based on PSLE AL scores (formerly T-scores), and competition for places is intense. The schools attract children from higher-socioeconomic-status families at rates that exceed their representation in the general population. ⚠Requires verification.

The International Baccalaureate is offered by several schools in Singapore, including some IP schools and some international schools. The IB Diploma Programme is a two-year course offered at the pre-university level (ages 16–19). Some Singapore students pursue the IB as an alternative to the A-Level system; its outcomes in terms of university admission are comparable. The IB's significance in the Singapore context is partly as an alternative assessment framework that emphasises different skills — extended essay, theory of knowledge, creativity-activity-service — and partly as a marker of differentiation for a specific socioeconomic segment of the student population.

The persistence of GEP and IP within a system otherwise moving toward greater flexibility is the most intellectually significant tension in the post-2024 education landscape. The government's stated position is that exceptional academic talent requires exceptional educational provision — that identifying and nurturing the highest-ability students is a social good because these students contribute disproportionately to national welfare through science, technology, public service, and intellectual production. Critics argue that the GEP and IP function as the preserved core of the elite reproduction system, ensuring that even if the mainstream structure is liberalised, the pipeline from wealthy family through elite primary school through IP secondary through PSC scholarship into the governing class remains intact.

---

11. Outcomes and Open Questions as of 2026

Singapore's education system in 2026 is structurally different from what it was in 2020, and radically different from what it was in 1985. The formal architecture of early-age sorting, stream labels, and T-score aggregate competition has been dismantled. The AL system has reduced fine-grained ranking anxiety at PSLE, at least at the margins. FSBB has removed the stream-based identity from secondary education. The GEP has been geographically dispersed. These are real changes, not merely cosmetic.

The outcomes that matter most are those that will not be visible for years. The 2024 Secondary One cohort under full FSBB will not complete their secondary education until 2028. Their post-secondary choices, polytechnic or junior college placements, and eventual labour market outcomes will not be measurable until well into the 2030s. The reforms have been designed carefully, but educational system reforms operate on long time horizons, and the evidence base for the specific design choices made in 2019–2024 is still accumulating.

Several open questions are particularly important. First: does FSBB's permeability actually function in practice? The theory is that students can move between G-levels within subjects as their performance evolves. The practice depends on school culture, teacher expectation, and the incentive structure for schools to support upward G-level movement. If the default is G-level stability — as stream placement was sticky in the old system — then FSBB's design advantage over the old system is unrealised.

Second: does the AL system's broadened bands actually reduce tuition demand, or merely redirect it? The concern is that if competitive pressure is redistributed — say, from marginal T-score improvement to ensuring AL1 rather than AL2 — the competitive dynamics and tuition demand simply reorganise around the new metric. Evidence from the 2022–2025 period is mixed on this question.

Third: what happens to the schools that were historically associated with Normal Technical streams? Under streaming, these schools had lower aggregate scores, lower teacher demand, and lower socioeconomic-status student bodies. Under FSBB, the same patterns will tend to reproduce through the admission ranking, unless active resource redistribution and school development policies interrupt the cycle. MOE's school partnership programme and resourcing commitments are relevant here, but their effectiveness needs sustained monitoring.

Fourth: does the Forward Singapore agenda — with its commitment to broadening definitions of success and recognising non-academic contributions — translate into measurable changes in employer hiring practices, scholarship selection, and public service promotion? The education system's incentive structure ultimately reflects what employers and institutions reward. If the credential hierarchy — A-Level, degree, prestigious university — remains paramount in labour market outcomes, the education system's internal reforms will be partially offset by external signals that re-emphasise credentials.

---

12. Conclusion

Singapore's journey from the EM1/EM2/EM3 streaming architecture of 1981 to the Full Subject-Based Banding system of 2024 is a case study in how a competent technocratic government manages the reform of a policy it cannot easily acknowledge was a mistake. The streaming system was built with genuine purpose — to address catastrophic educational wastage — and it achieved its aggregate objectives. It also created costs that were real, documented, and borne disproportionately by the most vulnerable children in the most vulnerable families. The government's response to this evidence has been to accept the diagnosis incrementally and to reform the system — EM3 abolition, T-score replacement, FSBB — while preserving the meritocratic legitimation narrative that sustained the system's political acceptance.

The result is a system that is genuinely less rigid, less labelling, and less consequential in its early-age sorting than the system that existed in 1985 or 2000. Whether it is equitable — whether it now provides genuinely equal educational opportunity to children from different socioeconomic backgrounds — is a different question, and the answer is less clearly positive. The private tuition industry, the elite school pathway, and the SAP school ethnic composition question remain as structural features of a competitive system that the label reforms have ameliorated without dismantling.

The intellectual debate — conducted by Teo You Yenn, Donald Low, Kenneth Paul Tan, and an emerging generation of Singapore-trained education researchers — has become more visible and more politically recognised than at any previous point. Forward Singapore's acknowledgement that Singapore had been defining success too narrowly was, in the context of these debates, a significant official concession. Whether it translates into the structural policy changes necessary to genuinely equalise educational outcomes is the central education policy question of Singapore's third generation of governance.

The streaming story is not over. The system built in 1981 has been substantially dismantled; the question of what replaces it — and whether the replacement can deliver both quality and equity — will occupy Singapore's education policymakers, teachers, parents, and students for decades to come.

---

Spiral Index

• For the foundational bilingual context that preceded streaming: SG-A-16 (The Bilingual Policy 1959–1979)
• For the full education policy arc 1959–2026: SG-D-02 (Education — From Colonial Classrooms to Global Rankings)
• For elite pathways, streaming, and social mobility in detail: SG-G-15 (Education System: Elite Pathways, Streaming, and Social Mobility)
• For the meritocracy ideology and stratification research: SG-J-07 (Singapore's Meritocracy: Promise, Reality, and the Stratification Research) and SG-M-02 (Meritocracy: Promise and Critics)
• For the architect of the 1979 Goh Report: SG-H-DPM-01 (Goh Keng Swee)
• For the eugenic dimension of education policy in the 1980s: SG-B-06 (The Graduate Mothers Scheme)
• For inequality trends and their education dimension: SG-O-08 (Inequality Trends)
• For the social contract framework within which education policy sits: SG-M-05 (The Social Contract)

---

Sources

1. Goh Keng Swee and the Education Study Team, Report on the Ministry of Education 1978 (Singapore: Ministry of Education, 1979).
2. Parliament of Singapore, Hansard, Select Committee of Supply debates on Education Ministry estimates, various sessions 1979–2026.
3. Ministry of Education, Singapore, "Refining PSLE Scoring and Secondary School Posting: Government Press Statement," September 2016.
4. Ministry of Education, Singapore, "Full Subject-Based Banding: MOE Press Release," November 2019.
5. Ministry of Education, Singapore, "Secondary Education Reforms — Implementation Updates," various press releases 2020–2024.
6. Ministry of Education, Singapore, Education Statistics Digest (Singapore: MOE, various years 2005–2024).
7. S. Gopinathan, "Education and the Nation-State: The Singapore Experience," in Education in Asia, ed. T. N. Postlethwaite (London: Routledge, 1980).
8. S. Gopinathan, ed., Education in Singapore: Emergence, Growth and Transformation (Singapore: Springer, 2022).
9. Jason Tan and S. Gopinathan, "Education Reform in Singapore: Toward Greater Creativity and Innovation?" NIRA Review (2000), pp. 5–10.
10. Jason Tan, Education in Singapore: Taking Stock, Looking Forward (Singapore: Prentice Hall, 2002).
11. Teo You Yenn, This Is What Inequality Looks Like (Singapore: Ethos Books, 2018).
12. Donald Low and Sudhir Thomas Vadaketh, Hard Choices: Challenging the Singapore Consensus (Singapore: NUS Press, 2014).
13. Kenneth Paul Tan, "Meritocracy and Elitism in a Global City: Ideological Shifts in Singapore," International Political Science Review 29, no. 1 (2008), pp. 7–27.
14. Tharman Shanmugaratnam, "Inequality and the Need for a New Social Compact," Amartya Sen Lecture, London, 2015.
15. Tharman Shanmugaratnam, Committee of Supply speeches on MOE estimates, 2003–2008, Parliament of Singapore Hansard.
16. Heng Swee Keat, "A More Flexible Education: MOE's Approach to PSLE Reform," National Institute of Education, 2012.
17. Chan Chun Sing, Committee of Supply speech on FSBB implementation, Parliament of Singapore Hansard, 2022.
18. Lee Hsien Loong, National Day Rally speech on education, 1997 ("Thinking Schools, Learning Nation").
19. Irene Y.H. Ng, "Education and Intergenerational Mobility in Singapore," Educational Review 66, no. 3 (2014), pp. 362–376.
20. OECD, PISA 2022 Results: Learning During — and From — Disruption (Paris: OECD Publishing, 2023).
21. Ministry of Education, Singapore, "Forward Singapore: Education Workgroup Recommendations," in Forward Singapore Report: Building Our Shared Future Together (Singapore: Government of Singapore, October 2023).