BAI: "Violation of Basic Principles of National Statistics Where Accuracy and Timeliness Are Essential"
"Undermines Policy Trust and Increases Future Housing Market Uncertainty"
Calls for Improvements to Unsold Housing Status Statistics and Automobile Management Information System
The Board of Audit and Inspection (BAI) has pointed out that the Ministry of Land, Infrastructure and Transport (MOLIT) identified a large-scale omission in the 2023 housing construction performance statistics but continued to calculate and publish the statistics without immediately releasing this information, potentially distorting the judgment of statistics users. The BAI stated that this violates the basic principles of national statistics, where accuracy and timeliness are essential, and demanded measures to prevent recurrence.
According to the audit report released by the BAI on January 15, the MOLIT became aware at the end of January 2024 that the 2023 housing construction performance statistics had been undercounted. After conducting its own review, the ministry only released the corrected figures on April 30, 2024. The scale of the omission was 40,000 units for permits (from 389,000 to 429,000 units), 33,000 units for construction starts (from 209,000 to 242,000 units), and 120,000 units for completions (from 316,000 to 436,000 units), totaling 192,330 units.
The housing construction performance statistics are a key housing supply indicator published monthly, following the flow of 'permit (approval) → construction start → completion.' As they are widely used as basic data for both government and local government housing policies, as well as for assessing the market's supply outlook, they are considered one of the major indicators.
In particular, the BAI found that even after the MOLIT became aware of the undercount, the ministry published the statistics for January and February 2024 without disclosing that the previous year's (2023) figures had been omitted. The statistical tables include comparison items such as the previous year's monthly performance and year-on-year change rates. If the previous year's figures are incorrect, the magnitude of increase or decrease is also inevitably distorted.
The BAI stated in the report that the MOLIT could have reduced user confusion by either obtaining approval from the National Data Office (formerly Statistics Korea) to remove year-on-year comparison items from the statistical tables, or by clearly indicating that "the previous year's figures are undercounted and are being corrected." However, the ministry failed to review or implement such measures. As a result, the BAI confirmed that the error in the year-on-year change rates in the statistics published for January and February 2024 ranged from a minimum of 10 percentage points (P) to a maximum of 88 percentage points.
In response, the MOLIT explained that it decided not to disclose the information, arguing that releasing it before the cause and scale were confirmed could lead to speculative media reports and increased instability in the housing market. However, the BAI concluded that this reasoning was unacceptable. The BAI stated, "According to the MOLIT's logic, it would be justified to conceal statistical errors on the grounds of potential market instability," and warned, "This undermines policy trust and poses a significant risk of increasing uncertainty in the future housing market." The BAI also noted that the MOLIT had already been briefed on the cause of the undercount, the estimated scale, and the correction period on February 15, 2024, and could confirm the final scale by the end of March. Therefore, at the very least, the ministry could have informed users about the possibility of errors and the correction schedule.
Why Were Over 190,000 Units Omitted? ... Code Omission During System Integration
The main cause of the omission was a 'code omission during the system integration process.' In line with the 2021 amendment of the Electronic Government Act, the MOLIT switched to receiving information from the building administration system 'Seumter' via the 'National Reference Data.' During this process, some business codes, such as those for urban redevelopment projects, were not included in the integration, resulting in their omission from the statistics, according to the BAI. The BAI stated that this 'code non-integration' accounted for 109,009 omitted units, representing 52.3% of the total omissions.
Another cause was 'delayed data entry.' The Housing Supply Statistics Information System (HIS) compiles statistics based on data finalized by city, county, and district governments for the previous month. However, since the system is based on ActiveX and became difficult to access after Internet Explorer support ended, local governments effectively stopped meeting deadlines. In fact, the closing rate for cities, counties, and districts plummeted from 95.7% in 2008 to 3.8% in 2023, and 35,503 units were omitted due to delayed data entry between January and June 2023.
Additionally, the BAI found that a 'code error' occurred when functions were changed during outsourced maintenance without consultation with the Korea Real Estate Board, resulting in 46,548 completed units being omitted. The BAI determined that the MOLIT and the Korea Real Estate Board had at least three opportunities to detect the omissions in advance-through data list checks, integration checks, and cross-checks with sales performance-but their oversight and supervision were inadequate.
Even After Corrected Announcement, Additional 1,176 Units Omitted
The BAI pointed out that even the corrected and republished 2023 statistics released by the MOLIT on April 30, 2024, still contained some omissions. Taking into account unreflected public institution housing construction performance (967 units), logical errors due to insufficient verification of source data (1,021 units), and duplicates (812 units), at least 1,176 units remained undercounted in the revised statistics. The BAI noted in a footnote to the report that the MOLIT further revised and republished the 2023 statistics on September 30, 2025, to improve accuracy, resulting in an increase of about 4,000 units compared to the April 30, 2024, figures.
This special audit also covered, in addition to the housing construction performance statistics, the unsold housing status report statistics and the automobile management information system. The BAI notified the MOLIT that although there have been ongoing concerns about undercounting in the unsold housing statistics, which rely on reports from developers, the ministry had failed to establish adequate accuracy verification measures and must address this issue.
Regarding the automobile management information system, the BAI pointed out that the MOLIT's approval and management procedures were lax, as it allowed the use of computerized data before the establishment of associations for online car auction companies and nonprofit organizations. As of September 2025, a total of 85 organizations (31 national agencies, 17 local governments, and 37 private entities) were receiving information via linkage with this system.
The BAI confirmed a total of eight violations or improper practices, including these matters, and demanded four cautions (involving three individuals), three notifications, and one notification (with corrective action completed). These special audit findings were finalized by the BAI Committee on December 18 of the previous year after internal review.
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.
