Japan has seen a number of soured public works projects now grappling with snowballing debts, ranging from toll expressways, gigantic bridges, airports and empty ports with huge container facilities.

A common feature of the failed projects was an optimistic demand projection made by bureaucrats that justified the initial master plans, only to turn out to be way off the mark when the facilities were opened after years of costly construction.

Some former senior government officials recently told The Japan Times that these projection "mistakes" were not accidental, at least for two very symbolic and notoriously debt-ridden projects: the Aqualine expressway connecting Tokyo and Chiba, and the Honshu-Shikoku bridges linking Shikoku with Honshu.

Though they knew from the outset the two projects would run in the red, the bureaucrats inflated their traffic demand forecasts to get the projects through the Finance Ministry's budget screening process before winning political approval to start the financially risky projects, they said.

"We cooked up (the traffic demand forecasts) to show the Finance Ministry. I think (bureaucrats) are still doing similar things today," Atsushi Shimokobe, who helped the former Construction Ministry conduct economic research for the Honshu-Shikoku project, said in a recent interview.

Shimokobe held key positions in the general development bureau of the Economic Planning Agency from 1962 to 1974 and is widely known as the mastermind of grand postwar national land development designs.

The Honshu Shikoku Bridge Authority, which manages the three routes, is now on the brink of bankruptcy with liabilities totaling 4.59 trillion yen -- much of which is expected to be eventually shouldered by taxpayers.

Meanwhile, with far less traffic volume than initial projections, the Aqualine toll expressway, part of which was constructed by the semigovernmental Japan Highway Public Corp., generated only 14.42 billion yen in revenues in fiscal 1999 while 45.87 billion yen was spent on interest and facility management.

The two projects have been under public scrutiny recently after a government panel tasked with privatizing four road-building public corporations pointed out their failures.

Panel members warn that similar manipulated traffic demand calculations for public road and bridge projects may occur in the future.

"An examination of the past cases is very important now, because the current demand projections could again lead to the construction of bridges like those to Shikoku," said Naoki Inose, a nonfiction writer and a key panel member.

In 1972, succeeding the joint feasibility study by the then Construction Ministry and Economic Planning Agency, the Honshu-Shikoku Bridge Authority estimated that in 1990, 37,900 automobiles would travel daily between Shikoku and Honshu on the Kobe-Naruto route, 25,500 vehicles on the Kojima-Sakaide leg and 15,300 on the Onomichi-Imabari segment.

But in 1990, only 8,321 vehicles a day traveled on the Onaruto Bridge, one of the two spans on the Kobe-Naruto route, and 9,809 traveled on the Kojima-Sakaide route. Unlike the 1972 estimate, the actual results also include automobiles that went to the middle segment of the toll expressways and returned without crossing the Seto Inland Sea.

In 1998, the Akashi Kaikyo Bridge, another component of the Kobe-Naruto route, finally opened. The advent of this bridge could only drive up the daily traffic on the Onaruto Bridge to 16,527 vehicles a day that year.

Shimokobe said bureaucrats are always required to submit forecasts to clear the Finance Ministry budget screening, no matter how groundless they may be.

"At best, an economic forecast is only valid for a period like 10 years," Shimokobe said. "So we usually start building roads without knowing what the actual traffic volume will be. I think this is the case with all road construction."

Political pressures

Kuniichiro Takahashi also admitted that he yielded to political pressure regarding the Honshu-Shikoku project. He was head of the Construction Ministry's road bureau when the Honshu-Shikoku Bridge Authority was established to construct all three segments of the project in 1970.

The ministry had long tried to adopt only one of the three proposed routes, believing construction of all three would not make economic sense given the huge costs and expected traffic volumes, Takahashi said.

But powerful politicians who were elected after promising their constituencies that bridges would be constructed pressured the ministry to adopt all three routes, Takahashi recalled.

"We had no choice but to 'lick the tip of the pencil,' " Takahashi said, meaning cooking up figures and documents.

When a political decision is made, bureaucrats have no choice but to follow it, even if they have to manipulate data to paint a rosy picture, he said.

"There's a limit to what bureaucrats can do," said Takahashi, who later served as vice construction minister and head of Japan Highway Public Corp.

Traffic demand forecast figures for the Aqualine were likewise concocted in the face of political pressure, according to Takahashi.

The Cabinet gave the green light to start construction of the gigantic undersea tunnel and bridge in 1985. Takahashi was head of Japan Highway Public Corp., which was later tasked with building part of the Aqualine.

"I repeatedly opposed (its construction), but it was no use," Takahashi said. "Since around that time, I knew (the Aqualine) wouldn't make ends meet."

The final forecast paper published by Japan Highway Public Corp. in 1987 concluded that Aqualine was viable, because 33,000 vehicles a day would take the toll road after its planned opening in 1996, and daily traffic would increase to 45,800 vehicles by 2001.

But the reality is way below that mark.

After taking eight years to construct, the Aqualine opened in December 1997. In the first year, daily traffic turned out to be a mere 11,900 vehicles, roughly a third of the initial forecast. It had increased to only 13,300 as of 2001.

Experts say long-term projections always have room for arbitrary manipulation, given the technical and other uncertainties over a long period.

Factors used to forecast traffic volumes usually include gross domestic product, population growth, expected regional development such as construction of industrial plants, and competition with other means of transport.

Long-term interest rate fluctuations, the growth rate of licensed drivers, the construction of other connecting roads and changes in the nation's industrial structure are also taken into account.

Tamotsu Okabe, who once headed the ports bureau at the former Transport Ministry, said bureaucrats have a tendency to choose favorable data for planned projects to clear the Finance Ministry's budget screening.

Rosy, 'serious' scenarios

Bureaucrats often prepare two different forecasts -- a "serious" one based on a realistic assumption and another one compiled based on optimistic figures, he said.

"There is a gap between serious (traffic projections) and those that were made up for budget (screening)," said Okabe, who led a number of harbor construction projects at the ministry. "You never use data that can be disadvantageous for the screening."

Okabe's remarks suggest bureaucrats by nature often try to expand their public works budgets regardless of the costs taxpayers and users will have to shoulder.

In its fiscal 1999 annual report, the Board of Audit pointed out that nine out of 14 regional airports whose comparable data were available saw smaller traffic than what was projected earlier. Four saw less than 50 percent of the initial forecasts.

"It's important to conduct a precise demand projection and thereby calculate the costs and benefits appropriately," the audit watchdog warned.

The Management and Coordination Agency pointed out in a 2000 report that 72 percent of 58 ordinary toll roads saw less traffic than initial Japan Highway Public Corp. projections.

In the face of recent criticism, the Land, Infrastructure and Transport Ministry last month revised its long-term traffic forecast for the entire country.

The new forecast, assuming a lower GDP growth rate than earlier forecast, projected that automobile traffic will peak in 2030 and began decreasing thereafter.

But writer Inose sees an apparent contradiction, pointing out that traffic volume growth is forecast to continue until 2030 while the nation's population is expected to peak as early as 2006.

At Wednesday's session of the government panel, Masatake Matsuda, chairman of East Japan Railway Co., also criticized the traffic projection.

Though the forecast took into account the rising number of elderly license holders, Matsuda argued that elderly drivers will not travel on expressways as frequently as younger, working people.