Seoul’s cram-school access shows up in housing prices, preprint finds

The price of private tutoring may extend beyond classroom fees. A Seoul housing study suggests families also commit more money to homes near university-entrance preparation academies—although it cannot establish that the academies themselves cause the higher prices.

Families seeking access to Seoul’s private cram schools may pay through housing as well as tuition. In a preprint posted on arXiv on November 4, 2026, Samiha Tariq reports that apartments near academies serving high-school students and university-entrance candidates command higher sale prices and lease deposits. The supplied metadata identifies a first-version preprint, not a peer-reviewed journal publication.

The paper’s central distinction is between paying for useful access now and paying because buyers expect that access to generate extra property appreciation. Its overall findings favor the first interpretation, subject to assumptions that prevent a definitive verdict on buyers’ expectations.

Two markets for similar homes

The analysis covers 2,197,510 apartment transactions in 8,427 Seoul complexes from January 2015 to October 12, 2026: 771,489 brokered sales and 1,426,021 jeonse contracts. That represents 93.2% of the eligible brokered sales and jeonse contracts citywide.

Jeonse tenants provide a large lump-sum deposit rather than monthly rent. The landlord must return the nominal deposit when the lease ends, although tenants face default risk. A larger deposit therefore ties up more household money; it is not equivalent to spending that entire amount on rent.

Tariq compares apartments within the same middle-school assignment zone and contract month, adjusting for characteristics including floor area, floor and building age. This helps separate tutoring access from public-school assignment, but does not equalize every school opportunity or neighborhood characteristic.

A premium concentrated around exam preparation

Using the official academy registry, the paper estimates that moving from a complex with six high-school or exam-preparation hagwons within 500 meters to one with 50 is associated with sale prices 11.8% higher and jeonse deposits 7.3% higher. Fifty academies is roughly the 97th percentile, making this a substantial access difference rather than a typical small neighborhood change.

Complex size changes the picture materially. Exam academies cluster near large apartment developments, which themselves sell at a premium. Holding complex size and builder characteristics fixed reduces the estimated sale premium to 5.6%, while the deposit estimate becomes 5.9%. Whether size is a confounding influence or partly bound up with the development of tutoring districts remains unresolved.

Other academic hagwons add no detectable sale premium once exam-preparation access is included. However, this comparison group includes unlabeled establishments, not just academies for younger children. More than half the registry records lack grade cues, and the two access measures are strongly correlated.

The sale-price finding withstands several alternative statistical checks. Evidence for the high-school-specific deposit premium is weaker: it loses conventional statistical significance with school-zone clustering and does not survive the paper’s full-family multiple-testing adjustments.

Access now, or appreciation later?

The paper’s strongest comparison groups sale and jeonse transactions by apartment complex, floor area rounded to one square meter, and month. These are matching flat types, not necessarily the identical dwelling: unit numbers are unavailable.

Across approximately 222,400 matched groups, exam-academy access produces no statistically detectable change in the deposit-to-sale-price ratio. For the six-to-50 comparison, the estimated ratio change is 0.4%, with a 95% confidence interval from −3.8% to 4.9%.

Tariq interprets this as consistent with access being reflected proportionally in both markets, rather than carrying an extra appreciation premium. Under the paper’s investor-landlord framework, and assuming no offsetting influences, the confidence interval allows expected appreciation to explain at most about 35% of the baseline sale premium. That conditional bound is not proof that appreciation expectations are absent.

The test detects substantially lower ratios in buildings at least 30 years old, consistent with reconstruction value. A possible exception also appears in newer buildings after mid-2021, although its within-complex result does not meet the stringent Bonferroni threshold.

Convenience, policy and remaining uncertainty

An initially larger lease premium using business-register academy counts disappears after accounting for nearby commerce. Shops and services raise deposits relative to sale prices even among matched flat types. The author cannot distinguish whether preferences, financing, perceived deposit safety or other mechanisms explain this pattern.

New tenants pay more than renewing tenants overall, but the study finds no lasting additional newcomer premium near exam academies. Education-policy announcement windows likewise show no clear net jump in the deposit premium; bundled policies and opposing effects limit that interpretation.

These remain conditional associations, not causal estimates of opening a hagwon. Affluent households may attract academies; school quality is incompletely observed; historical school boundaries are unavailable; and the 2026 academy registry omits closed establishments. Location estimates also have limited independent validation.

The supplied extraction includes appendices and references, but some figure labels and table layouts are garbled, so the numerical account here relies on explicit prose and readable tables. The paper identifies a potential housing component of educational competition—not demonstrated effects on exam success or fertility.