Busan Is Sorting Its Students Before They Reach the Classroom

Busan’s school divide is often framed as a problem for the education office: overcrowded classes in new towns, shrinking schools in older districts, and a superintendent expected to close the gap through construction plans, welfare programs, AI tools and teacher protection. The framing became more visible after Kim Seok-joon returned to the Busan education office for a fourth term. It also narrows the story too quickly.

The divide begins before a child reaches a classroom. Busan’s housing market, apartment-led development pattern and uneven neighborhood decline are reshaping the school system before education policy can respond. Young households are concentrating in new apartment districts such as Myeongji, Eco Delta City, Jeonggwan and Ilgwang, while older districts lose children, schools and residential diversity. Classroom pressure, in that sense, is not merely a matter of school capacity. Busan’s education map is increasingly following its housing map.

An analysis of 2026 public school disclosure data covering 615 elementary, middle and high schools in Busan shows 278,026 students across 13,057 classes, with an average class size of 21.3 students. The average suggests manageable pressure. The distribution tells a different story. Eighty-five schools have at least one grade cohort with an average of 28 or more students per class, while 52 schools have fewer than 100 students in total. The same city contains schools where particular grades are too crowded for comfortable instruction and schools where the entire student body has fallen below the scale of a single large apartment cohort.

Kim’s fourth term will be measured against familiar promises: AI education, universal education welfare, regional equity and stronger protection for teachers. Those pledges matter because they shape budgets, staffing and daily school operations. Yet the deeper test lies outside any campaign platform. Busan is losing students unevenly, concentrating many of its remaining young families into a limited number of residential districts and leaving older neighborhoods to manage decline. The city is sorting students before schools ever teach them.

The superintendent frame is too narrow

Kim Seok-joon’s return gives the story an obvious political opening. A fourth-term superintendent carries unusual responsibility, and Busan’s education agenda now includes regional learning gaps, overcrowded schools, small-school decline, student mental health, teacher protection and the transition to AI-supported learning.

The education office can act on many of these pressures. It can adjust student placement, support small schools, revise school zones, expand welfare, request school construction, deploy digital tools and strengthen legal or administrative support for teachers. These are not symbolic measures. They affect classrooms, timetables, families and teachers.

Administrative authority, however, does not reach the full origin of the problem. Busan’s school divide is shaped by where apartments are built, which households can afford them, when families move in, how older neighborhoods lose young residents, how long school approval takes, and how transport, care, local services and private education markets influence family decisions. By the time overcrowding appears in a classroom, the deeper decisions have already been made through housing markets, development schedules and urban planning processes.

A superintendent-centered frame assigns too much responsibility to one office and too little attention to the city producing the pressure. New-town parents demand more classrooms and safer school access. Older districts fear closures, mergers and the loss of neighborhood anchors. Teachers face larger administrative burdens as schools absorb both crowding and decline. The education office becomes the agency expected to repair the consequences of urban development after those consequences have already reached children.

Kim’s fourth term should therefore not be judged only by whether the number of overcrowded classes falls or whether AI education expands across schools. Those goals are measurable and politically visible, but they do not define the full scale of the challenge. Busan needs to stop treating school pressure as a series of local emergencies and start reading it as a citywide pattern of residential sorting.

The limit of the education-office frame becomes clearest in the contrast between new towns and older districts. In Myeongji, Eco Delta City, Jeonggwan and Ilgwang, school demand is shaped by large apartment projects and the arrival of young families in concentrated cohorts. In older districts, declining enrollment reflects aging, housing deterioration, youth outflow and weakened neighborhood reproduction. The education office can respond to both conditions. It cannot alone control the forces that create them.

A serious education agenda for Busan must connect student placement, school construction, small-school policy, AI investment and education welfare to the city’s housing map. Schools are not separate from urban planning. They are one of the clearest measures of whether a city can still hold different generations, income groups and family types in a shared civic fabric.

The average hides the geography of stress

Busan’s school system appears calmer when reduced to a citywide average. The 2026 disclosure data analyzed for this article covers 278,026 students in 615 elementary, middle and high schools, divided into 13,057 classes. The average class size is 21.3 students. On its own, the figure suggests a system under manageable pressure.

Stress appears instead in specific schools, grades and residential districts. In the same city where the average class has just over 21 students, 85 schools have at least one grade cohort with an average of 28 or more students per class. At the other end of the system, 52 schools have fewer than 100 students in total. Those indicators belong in the same analysis because they describe the same urban process: the concentration of children in some neighborhoods and the withdrawal of children from others.

The contrast sharpens when schools are grouped by residential pattern rather than administrative district. The current overcrowded new-town cluster has 26,697 students in 28 schools, or 953 students per school. The old-district and small-school cluster has 23,086 students across 74 schools, or 312 students per school. The industrial and outer-area cluster has 10,664 students in 39 schools, or 273 students per school. A citywide average cannot capture a school system in which one set of schools absorbs large apartment-based cohorts while another struggles to maintain institutional scale.

District averages flatten the problem in a similar way. Gangseo District is often described through Myeongji and Eco Delta City, the symbols of western Busan’s development frontier. The same district also contains outer and semi-rural school zones where enrollment remains thin. Gangseo has 21,904 students across 39 schools, yet it also has 11 small schools and seven schools showing decline signals. Growth and hollowing sit inside the same administrative boundary.

Gijang County shows the eastern version of the same pressure. Jeonggwan and Ilgwang have pushed school demand upward through the education system. Elementary demand in a new residential district becomes middle-school pressure and then high-school pressure. Gijang has 20,607 students across 35 schools, but the pressure is not evenly distributed. Large secondary schools such as Jeonggwan High School and Sinjeong High School show how apartment cohorts move through the system, forcing the city to respond level by level, often after the pressure has already arrived.

The year-over-year pattern makes the problem harder to dismiss as ordinary demographic decline. From 2025 to 2026, the growth-prospective new-town cluster gained 1,716 students. During the same period, the old-district and small-school cluster lost 962 students, the industrial and outer-area cluster lost 780, and the first-generation planned-city cluster lost 635. Busan is not experiencing a simple, uniform fall in student numbers. The city is experiencing spatial redistribution.

A city can lose students overall and still produce overcrowded schools if young families concentrate in fewer residential districts. A city can report a moderate average class size and still leave some schools with crowded grade cohorts, strained facilities and compressed teacher attention. A district can contain new apartment growth and small-school decline at the same time when development is concentrated in one part of the district while older settlements lose children.

The usual administrative categories are too narrow. Overcrowded classes, small schools, school closures, relocations, modular classrooms and school-zone adjustments are often treated as separate policy problems. The data point to a larger pattern: Busan’s school system is being reshaped by the uneven movement of families across the city.

The core finding is not that some schools are crowded and others are shrinking. Both conditions are being produced together. New-town schools are absorbing synchronized apartment cohorts. Older and outer districts are losing the children that once sustained neighborhood schools. A divided education map is emerging from a divided housing map.

New-town growth is internal sorting, not recovery

The growth of school demand in Busan’s new towns can look like demographic recovery. Myeongji, Eco Delta City, Jeonggwan and Ilgwang have the visible signs of a younger city: new apartment towers, new roads, new schools, new commercial strips and parents demanding more classrooms. Those districts appear to contradict the broader story of a shrinking and aging Busan.

The appearance can mislead. New-town growth does not necessarily mean Busan as a whole is becoming younger. The pattern points instead to a concentration of young families into fewer residential choices. A city can produce new-town expansion without producing urban renewal. Children can move from one part of the city to another while the total pool of children continues to decline.

School geography registers that movement quickly. New residential districts absorb families with children, while older districts lose the children needed to sustain neighborhood schools. The result is not a balanced redistribution of opportunity. It is a narrowing of the places where family life appears viable.

The language of “new-town demand” can turn internal redistribution into a false story of recovery. Growth can sound external, as if young households are entering Busan from elsewhere and adding to the city’s future. Some inflow from nearby cities or other regions may exist. The larger urban question remains internal. When families leave older neighborhoods for new apartment districts inside the same metropolitan system, the city relocates its demographic problem rather than resolving it.

Schools register residential relocation faster than many other institutions because every family move appears quickly as a class-size problem, a school-zone dispute, a school-bus demand or a closure debate. A young family moving into a new apartment district affects kindergartens, elementary grades and after-school demand almost immediately. A few years later, the same cohort moves into middle schools. Later, the pressure reaches high schools. The city experiences each stage as a separate school-capacity emergency, while the source lies in a residential pattern set years earlier.

Gangseo shows why the growth narrative fails. Myeongji and Eco Delta City symbolize western Busan’s expansion, but Gangseo also contains outer settlements and school zones where enrollment remains thin. Growth and decline are not separated by district boundaries. A single district can contain a crowded new-town school and a small school struggling to maintain scale.

Gijang shows the same logic in another form. Jeonggwan and Ilgwang are not only residential expansion areas. They are cohort machines. They concentrate households at similar life stages, then push school demand upward through the education system. Elementary pressure reappears as middle-school and high-school pressure. Each stage produces a new administrative response: classroom additions, school-zone adjustments, modular facilities, temporary campuses or relocation debates.

Older districts experience the reverse process. They lose students not only because birth rates fall, but also because family formation, housing preference and educational expectation move elsewhere. A smaller school can mean fewer classes, fewer peers, fewer programs, weaker extracurricular life and less confidence among prospective families. The neighborhood then becomes even less attractive to households with children.

The feedback loop carries long-term danger. New towns receive students and the social confidence that older neighborhoods are losing. Parents move toward places where other parents have already moved. Schools grow where school demand is already visible. Academies and commercial services follow the same map. Housing prices absorb the educational premium and filter future residents more tightly. The education divide becomes self-reinforcing.

New-town growth should not be celebrated too quickly as a sign of urban vitality. The pattern may indicate concentration rather than renewal. Young households may still see a future in Busan only when that future is packaged as a new apartment district with predictable schools, parking, roads, retail and private education access. That is a narrow model of family life for a metropolitan city.

The cost appears on both sides. New-town schools face crowded grade cohorts, strained facilities and reduced individual attention. Older schools face declining enrollment, weaker peer networks and the loss of neighborhood function. One side is pressured by too many students moving together. The other is weakened by too few students remaining. Both forms damage educational quality.

School construction in growth districts remains necessary, but construction alone cannot explain why the same pressure keeps returning. The deeper question is why Busan keeps producing growth districts that generate school-demand surges while older neighborhoods lose the conditions needed to hold families.

Apartment urbanism changed the school itself

New-town school overcrowding is often treated as a failure of supply. More students arrive than expected, classrooms fill, parents demand relief, and the education office responds with new schools, additional classrooms, modular buildings or temporary campuses. The sequence is familiar enough to appear administrative.

Apartment-led development creates the rhythm that overcrowds schools. Large apartment districts do not bring families into a city gradually. They concentrate them. A new residential district is usually occupied within a compressed period by households sharing similar life stages, housing preferences, income thresholds and educational expectations. The district does not simply gain residents. It gains cohorts.

The first pressure may appear in childcare and early elementary grades. A few years later, the same pressure moves into upper elementary grades and middle schools. Later, high schools absorb the cohort. The city experiences each stage as a separate school-capacity problem, although the same residential cohort is moving through the system.

A new-town school may become overcrowded not because the city is growing sustainably, but because a large group of similar households entered the same district at the same time. The school is forced to absorb the timing of the housing market.

The timing creates a policy trap. If the education office builds too little, students spend years in crowded classes, temporary facilities or unstable school-zone arrangements. If it builds too much, the same district may face underused classrooms once the cohort moves on. The dilemma is built into synchronized apartment settlement.

Modular classrooms and temporary second campuses reveal the dilemma. Some districts may need temporary capacity because permanent construction can overbuild for a peak that may not last. Temporary capacity, however, also exposes the weakness of the system. The city has produced a residential district whose school demand is too intense to ignore and too uncertain to answer with permanent capacity alone.

Overcrowding affects education beyond the number of seats. Crowded schools reduce teacher attention, stretch counseling capacity, strain lunch schedules, compress special-room use, crowd playgrounds and libraries, and make school activities harder to organize. Class size is only the most visible measure of a wider institutional burden.

Additional buildings cannot solve every part of that burden. School capacity is not only a matter of classrooms. A school needs teachers, subject diversity, stable peer groups, safe routes, sports space, libraries, counseling, after-school care and a relationship with its neighborhood. When schools are planned as late responses to apartment occupancy, education becomes an accessory to housing rather than a condition of urban life.

Apartment districts also change the social character of schooling. A large new complex may contain many children, but scale does not guarantee diversity. Similar housing prices, tenure structures, move-in dates and family stages can create a socially narrower school environment. Students may be numerous, while the range of household backgrounds becomes thinner than in older mixed neighborhoods.

The paradox matters. New-town schools can be crowded and homogeneous at the same time. They may have the energy of a young district, but they also concentrate similar parental anxieties, similar educational expectations and similar forms of competition. The school becomes less a meeting place for a varied neighborhood and more an extension of a filtered residential market.

Older Busan tells a different story. The city’s past should not be romanticized. Older neighborhoods contained hillside poverty, overcrowded homes, port labor districts, uneven school conditions and clear differences between neighborhoods. Many children grew up in difficult urban environments long before today’s new towns appeared.

The older city, however, was unequal in a different way. Its neighborhoods were often more socially mixed because the city itself was more compressed. The port, markets, factories, hillside settlements, commercial streets, small houses, rented rooms and public institutions sat close to one another. Families with different jobs, incomes and migration histories often lived within overlapping school catchments. A neighborhood school could draw children from shopkeeper families, port workers, factory employees, civil servants, tenants, small landlords, street vendors and households formed by war, labor migration or industrial expansion.

The mixture did not erase inequality. It often made inequality visible inside the same classroom. The school still functioned as one of the few institutions where a varied neighborhood could be gathered under a shared public routine. The old school was not necessarily better equipped. It was not always fairer. It was more likely to reflect the social mixture of a dense urban district.

Busan’s urban history made that mixture unusually intense. The city absorbed refugees during and after the Korean War, expanded around the port and manufacturing economy, and grew through hillside settlements and working-class neighborhoods built under pressure rather than design. Sanbokdoro communities, old commercial districts and port-adjacent neighborhoods carried poverty and infrastructural weakness, but they also placed different forms of urban life near one another. Schools inherited that complexity.

The shift toward planned apartment districts changed the relationship between neighborhood and school. Haeundae New Town, Hwamyeong and later Myeongji, Jeonggwan, Ilgwang and Eco Delta City did not simply provide new housing. They introduced a different way of organizing families, schools and social expectations. The apartment complex became not only a residential form but a sorting device.

Sorting occurs through price, tenure, unit size, move-in timing, school reputation, commuting patterns, parental expectations and access to private education. A large apartment district may contain thousands of residents, but if the housing product is similar, the price band is similar, the move-in date is compressed and the target household is a family with children, school population can become large while social range narrows.

Older Busan produced difficult mixtures. Newer apartment districts produce cleaner separations.

The contrast changes what kind of public institution a school becomes. A school shaped by mixed streets and overlapping livelihoods absorbs the irregularity of an urban neighborhood. A school shaped by synchronized apartment cohorts absorbs the pressure of a residential product. In the first, inequality appears through differences inside a shared district. In the second, inequality is increasingly sorted before the child arrives at school.

Old districts do not need nostalgia. They need repair: better housing, safer streets, stronger care infrastructure, improved schools and serious public investment. Still, the old city’s mixture contained a civic possibility that current policy should not discard. Children from different household types and social backgrounds could share a local school because their families still shared the same neighborhood.

That possibility is weakening. As young families move toward new apartment districts, older schools lose enrollment, peer networks and program scale. New-town schools gain students in large, synchronized cohorts. The city is not only changing the number of students in each school. It is changing the social composition of schooling. Housing now sorts children before education policy reaches them.

Busan needs a city-education plan

Busan does not need only an overcrowded-classroom reduction plan. Busan needs a city-education plan.

An overcrowding plan begins after pressure appears. It counts students, adds classrooms, adjusts school zones, opens modular facilities, relocates schools or requests new construction. Some of those measures are necessary. Some are urgent. They treat the school as the site of the problem, rather than the place where a larger urban problem becomes visible.

A city-education plan would begin earlier. It would place school capacity inside the same planning frame as housing approval, apartment move-in schedules, student generation, safe routes, public transport, care infrastructure and old-district regeneration. Busan should be able to see school pressure forming before it reaches the classroom, not after parents begin demanding emergency relief.

The first task is a living school-capacity map. District averages cannot explain Gangseo, where Myeongji and Eco Delta City coexist with outer school zones losing scale. They cannot explain Gijang, where Jeonggwan and Ilgwang produce secondary-school pressure while other areas remain thinner. Busan needs neighborhood-level data that combines school enrollment, grade-level overcrowding, small-school signals, apartment occupancy, housing type, student projections, walking routes, transit access and after-school care.

The second task is a Busan-specific student generation model. Large apartment districts produce different school demand depending on unit size, price, tenure, public-rental share, move-in timing, household age, school reputation, academy access and commuting patterns. Underestimating student generation pushes children into overcrowded classrooms and unstable assignment. Overestimating demand builds permanent capacity for a temporary cohort peak. Both errors become more avoidable when student projection becomes a serious urban-planning tool rather than a late education-office calculation.

The third task is to move school planning to the housing-approval stage. Schools should not be public facilities attached after apartments are sold. Before large residential projects are approved, Busan should examine school sites, grade-level demand, middle- and high-school capacity, safe routes, lunch facilities, sports space, libraries, care programs and transit links. The relevant question is not whether a site can fit housing units. The relevant question is whether the district can sustain family life without transferring the cost of poor planning to children.

The housing model also has to change. Apartments are central to Korean urban life, and families choose them for rational reasons: safety, maintenance, parking, transport, retail, school access and predictability. The problem is not the apartment itself. The problem is a development model that turns the apartment complex into the dominant container of family life. If every growth district is built around similar apartment products, similar price bands and synchronized move-in cohorts, Busan will keep producing school-demand waves. Stable schools require neighborhoods where different household types, income levels, tenures and life stages can coexist.

Old-district schools require a different policy language. Some small schools may need consolidation, and no city can maintain every school indefinitely if enrollment collapses. Closure, however, should not become the default language of decline. In some neighborhoods, a small school may be the last institution capable of anchoring family life. A school can become a site for care, libraries, youth programs, lifelong learning, cultural activity and community services. If older districts are expected to attract young households again, schools must become part of their recovery rather than evidence of their decline.

AI education belongs inside this geography, not above it. Digital tools may help diagnose learning gaps, support writing, personalize practice or supplement small-school instruction. They cannot substitute for a school network that is spatially unbalanced. In overcrowded schools, AI does not create more teachers, counseling time, libraries, lunch capacity or safe commutes. In shrinking schools, AI does not restore peer networks, course diversity or neighborhood confidence. Technology spending should follow the geography of inequality rather than float above it as a campaign slogan.

The governance problem is as important as the policy design. Housing is planned by one system. Schools are managed by another. District governments handle local services. Developers follow project incentives. Families experience all of these as one life decision. Busan needs a formal structure linking the city government, the education office, district governments and development agencies around shared school-and-housing indicators. Without that structure, each institution can claim to be doing its job while the city continues to reproduce the same imbalance.

A four-year term rewards visible measures: new classrooms, school openings, AI devices, welfare programs and budget announcements. A city-education plan works on a longer timeline. It asks whether Busan can maintain mixed neighborhoods, stable schools and family life across more than a handful of new residential districts.

Busan’s education divide is a warning about the city’s future. A city that concentrates young families in standardized new towns while older districts lose children is not solving its demographic crisis. It is reorganizing it. Building more schools may be necessary, reducing crowded classes may be urgent, and supporting small schools may be humane. None of those measures will be enough if Busan keeps producing the same urban structure underneath them.

The next stage of education policy should begin outside the classroom, with the housing map, the age map, the school map and the question of whether different generations and incomes can still share the same city.

Busan is sorting its students before they reach the classroom. Public policy now has to decide whether to manage the consequences or confront the city that produces them.