Police Search Toyama City Hall Over Alleged Census Inflation

A preliminary population of 402,133 became a final 399,197. The ministry’s criminal complaint and a police search now put the integrity of one of Japan’s most important statistical systems under scrutiny.

TOYAMA. Japan’s national census is more than a head count. Its figures feed government finance, urban planning, representation, research and the statistical models used to understand the country. That is why an alleged manipulation of a few thousand people in one city has escalated into a criminal investigation.

Toyama Prefectural Police searched Toyama City Hall on October 3 over suspected violations of the Statistics Act connected to the 2025 census. About 10 investigators arrived at the municipal building around 10 a.m. and went to the fifth floor, where the division responsible for census work is located, according to Jiji Press.

Four days earlier, on September 29, the Ministry of Internal Affairs and Communications filed a criminal complaint with prefectural police against unidentified Toyama city employees. The ministry alleges that the city’s census population was improperly inflated by at least 2,000 people and that multiple staff members were involved.

The official final population of Toyama City in the 2025 census is 399,197. The preliminary figure published in May had been 402,133—a downward revision of 2,936.

What has not yet been established publicly is who changed which records, whether anyone ordered the changes, and what motive—if any—lay behind them. The case is under investigation. No criminal guilt should be inferred from the search or complaint alone.

From 402,133 to 399,197

Japan’s census counts people according to where they usually live as of October 1 every five years. It is not simply a copy of the resident registry, and differences between census population and resident-registry population are normal.

Toyama City’s resident-registry population at the end of October 2025 was 401,797. The fact that the final census count was lower—399,197—does not itself indicate wrongdoing. Students living away from home, people who have moved without immediately changing registration and other differences in residence definitions can produce legitimate gaps.

The issue is the discrepancy between the census’s own preliminary and final counts, combined with what the ministry says it found during review of the data.

The preliminary population published in May was above 400,000. The final figure fell below that line. At the prefectural level, the 2025 census found 982,147 residents in Toyama Prefecture, down 52,667, or 5.1%, from 2020.

Why a few thousand people can affect public money

Census numbers are embedded in Japan’s fiscal and administrative systems. The Statistics Bureau has long described census population as a basis for calculations including local allocation tax grants and subsidies, as well as for representation, planning and sample surveys.

Local allocation tax is a national transfer designed to reduce differences in fiscal capacity among municipalities and help provide a broadly comparable level of public services across Japan. Its calculation uses multiple demographic and fiscal measures, so population can affect the amount a municipality receives.

The fiscal 2026 calculation used the preliminary census population. Toyama City has estimated that it may have received about ¥60 million more than it was entitled to because of the higher preliminary count.

That figure is a city estimate, not yet a final repayment determination. How the national government will reconcile the overpayment, if confirmed, remains to be settled.

Mayor Hirohisa Fujii denies an order from senior officials

Toyama’s mayor is Hirohisa Fujii, who was first elected in 2021 and reelected in April 2025.

At a news conference on October 2, Fujii denied that the alleged inflation was intended to keep the city’s population above 400,000 and said there had been no instruction from senior officials to inflate the figure, according to Jiji. He said he would take appropriate responsibility at the proper time.

After the October 3 search, the city said it was taking the investigation seriously and would continue to cooperate fully.

The distinction matters. The preliminary count being above 400,000 and the final count being below it are facts. A motive to preserve that threshold has not been established by investigators, and the mayor has denied it. The investigation must determine who did what and why.

The prefecture says trust in the survey itself has been damaged

On September 29, Toyama Prefecture issued an official statement after the ministry’s criminal complaint. The prefecture said the case undermined not only trust in the results but trust in the census itself among residents, data users and people who worked on the survey.

That is a serious institutional concern. Japan’s Statistics Act imposes confidentiality duties on census workers and criminal penalties for breaches because the system depends on households answering sensitive questions in confidence.

Accuracy is the other half of that compact. People must believe not only that their answers will remain protected, but that public officials will aggregate them faithfully.

A census dating to 1920

Japan conducted its first modern national census on October 1, 1920. The Statistics Bureau describes it as the beginning of the country’s modern statistical system. That census counted 55,963,053 people.

The project was politically significant. Prime Minister Hara Takashi’s government created the Census Bureau as part of a broader effort to build the statistical capacity of a modern state. Publicity for the first census included lectures, parades and nationwide campaigns, and the exercise was treated as a major national event.

With wartime exceptions, Japan has continued conducting the census every five years. Over the following century, the survey became the baseline for population estimates and a central source for information about households, housing, work, commuting and migration.

The Statistics Bureau has described census data as a “shared asset of the people” and an information foundation for society. That long institutional history is why alleged manipulation in a municipal processing office has consequences far beyond one city’s population ranking.

Why preliminary and final numbers can legitimately differ

A preliminary census count is designed to provide a fast population and household measure before every questionnaire and consistency check has been fully processed. The final basic tabulation follows deeper review and cleaning.

Some movement between preliminary and final results is normal. Duplicate records, incomplete questionnaires, residence classification and household handling can all require adjustment.

What makes Toyama different is that the internal affairs ministry did not treat the full revision as ordinary statistical cleaning. It filed a criminal complaint after concluding that at least 2,000 people had been improperly added and that multiple employees may have been involved.

The precise method allegedly used has not yet been made public. Nor has the ministry publicly broken down how much of the 2,936-person revision was ordinary correction and how much is part of the suspected misconduct.

The resident registry is not the census

Confusion between the two population measures can obscure the case.

The resident registry is an administrative record updated continuously as residents register births, deaths, moves and other changes. Municipalities use it for day-to-day administrative purposes. Since 2012, registered foreign residents are included as well.

The census, by contrast, is a statistical survey conducted every five years and counts people according to usual residence on the census date.

Toyama City’s resident-registry population was 401,797 at the end of October 2025. By the end of August 2026 it had fallen to 399,759. The city was therefore moving naturally around the 400,000 level even in its administrative population records.

Population decline makes precision more important, not less

Japan’s national population continues to shrink. The Statistics Bureau estimated the September 2026 population at about 122.65 million, down roughly 540,000 from a year earlier.

Toyama Prefecture’s final 2025 census population was 982,147, 5.1% below the 2020 count. That decline affects decisions about schools, hospitals, public transportation, utilities, emergency services and social care.

In a growing city, a planning error may be absorbed by expansion. In a shrinking city, small differences can influence when a school is consolidated, how frequent a bus service should be, where health capacity is placed or how much infrastructure a city can afford to maintain.

Toyama is particularly notable because it has become internationally associated with compact-city planning and data-driven urban policy. The city publishes fine-grained population data on a 250-meter grid derived from resident-registry records to monitor urban development. A census-integrity investigation therefore cuts directly across a city that has made population data central to its planning identity.

A criminal case and an administrative failure are not the same inquiry

The police investigation will focus on evidence of offenses under the Statistics Act or other laws and on individual criminal responsibility.

Public administration requires a separate review. Who could edit or supplement census records? Were changes logged? What supervisory approval was required? What anomaly checks existed? Why was the problem not caught before publication of the preliminary figure?

Even if investigators ultimately conclude that a small number of employees acted independently, the ability to alter enough records to affect a major national statistic would still raise an internal-control question. Conversely, if some disputed adjustments turn out to have been legitimate statistical processing, those procedures should be explained clearly rather than treated as misconduct.

Five questions still unanswered

First, what exact data operation led the ministry to characterize the count as improperly inflated? Second, how many employees were involved and what authority did they have? Third, did supervisors or senior officials know of the adjustments? Fourth, how will the estimated local-allocation-tax overpayment be reconciled? Fifth, do census systems nationwide need stronger audit trails and anomaly detection to prevent a recurrence?

None of those questions is resolved merely by the police search.

The final population of 399,197 is now Japan’s official census figure for Toyama City. The next task is to explain, with documentary evidence, why the preliminary figure was 2,936 higher and to distinguish normal census correction from the conduct that triggered a criminal complaint.

Statistical trust does not require a system that never discovers an error. It requires a system that can expose the error, correct the number, establish responsibility and show how the same failure will be prevented next time.