Systematic miscounting may have cost the Hong Kong government hundreds of millions of dollars in unpaid tolls over the last 10 years

An exposé by financial analyst and activist investor David Webb has found serious miscounting of cross-harbour tunnel traffic over the last decade.

According to Webb, government traffic figures showed Hong Kong Island to be a net vehicle exporter, with over 1.2 million vehicles apparently shifting from Hong Kong Island to Kowloon in 2022 alone.

With no vehicle manufacturing on Hong Kong Island and only a small number of vehicles able to leave the island by other means, Webb says the “uni-directional flow” likely reveals systematic undercounting in the southbound direction of the three harbour tunnels.

The phenomenon was seen across all vehicle types, from double-decker buses to motorcycles.

As drivers often use different harbour tunnels for their north and southbound journeys, the net northbound/southbound flows at each individual tunnel would not be expected to cancel out, but the aggregate “three tunnel” net vehicle flow between Hong Kong Island and Kowloon should roughly balance over time – as indeed they now do under the HKeToll system, introduced in 2023.

“The discrepancies have almost vanished since eToll came in, so any undercounting (or overcounting) is [now] consistent in both directions,” says Webb.

Explanations such as new vehicle registrations or Hong Kong’s very limited vehicular ferry services don’t hold water: the net northbound flows reach many multiples of the total vehicle fleet.

For example, according to Transport Department (TD) data collated by Webb, around half a million double-decker buses were “exported” from Hong Kong Island between 2013 to 2022.

“We can assume that they all had Autotoll tags installed, but there is still a massive discrepancy per year which is far larger than the licensed bus fleet of about 5,000. So this does suggest a defect in the Autotoll system of some kind?” Webb asked in one of a series of emails to TD.

But officials have so far refused to engage on the issue, shrugging off Webb’s questions and claiming repeatedly that the tunnel figures are “audited” by a professional accountant.

Officials gave Transit Jam the same response.

“The traffic figures captured before the implementation of HKeToll had been audited and accepted as appropriate by the Certified Public Accountant (Practising) engaged by the respective tunnel operators,” a TD spokesperson said, in a line also parroted to Webb at least twice since December 2023.

TD refused to share the name of the auditor and would not share the audit report.

But Webb, who runs the Webb-site corporate governance portal, says the audit is irrelevant and does not explain the mismatch in northbound and southbound traffic.

The issue is not just academic. “[R]emember that these data have been relied upon to reach policy decisions such as the recent rebalancing of fares in the three tunnels,” he wrote on Webb-site.

And the lost revenue is significant. Taking data from 2022, the last full year before HKeToll was introduced, Hong Kong Island “exported” around 78,000 buses, 192,000 trucks, 913,000 private cars, 26,000 taxis, 7,000 minibuses and 30,000 motorcycles through the three cross-harbour tunnels.

Assuming these numbers arise from under-counted southbound trips, and even assuming all those uncounted trips occurred at the lowest-tolled tunnel (the Cross Harbour Tunnel), the missing tolls add up to HK$23 million of missing revenue.

While this is just 1% of the government’s estimated HK$2 billion annual citywide bridge and tunnel income (before it took ownership of the Western Harbour Tunnel), it’s around 30% of Cross-Harbour Tunnel annual revenues; and as the miscounting appears to have been going on for at least a decade, the government coffers could be shy of several hundreds of millions of dollars due to the error.

Nobody has a clear explanation for the phenomenon. Before HKeToll, vehicles were counted either as they paid cash tolls or through the AutoToll system. If the AutoToll system failed to register an AutoToll tag for any reason, the AutoToll system would take videos of the vehicle passing through “for the tunnel operators’ follow-up actions,” according to the government.

“The tunnel operators would then review the toll evasion cases manually, and follow up with the Autotoll company or vehicle owner on the outstanding toll as appropriate,” the government said in an email to Webb in April 2024.

It’s most likely this “manual follow-up” is the failure point.

This could be due to the technical challenges of identifying licence plates from fast-moving vehicles in one direction or other – the toll booths for all three cross-harbour tunnels are on the Kowloon side, meaning licence plate capture for non-paying northbound and southbound vehicles will be from different directions. From the historical placement of cameras and booths at Cross-Harbour Tunnel, for example, it’s likely errant southbound vehicles were photographed from the rear while northbound vehicles were snapped from the front. If rear plates (or front plates, if the photographing is the other way around) are statistically more likely to be obscured or covered or somehow harder to read then this would explain the undercounting.

Or the issue could be tunnel staff simply not bothering to follow-up small infringements; the payoff in chasing a motorcycle rider to recover HK$8 or taking a bus firm to court for HK$15 may have been seen as a fool’s errand, either at a corporate level or by individual staff.

And corruption is always a possibility in Hong Kong, given the number of Autotoll buses escaping the toll. Yet it’s hard to imagine what would be a fairly complex decade-long conspiracy for what would yield the bus companies collectively just a HK$1 million saving a year.

Whatever the explanation, the fact remains that millions of vehicles had likely been riding southbound tunnels for free for years.

The government says it had instigated random video checks at the Cross-Harbour Tunnel and Eastern Tunnel to double-check tunnel operator figures. “During our random inspections, no major discrepancies in the traffic figures have been found,” it told Webb in April.

While it’s interesting the Hong Kong government doesn’t view a miscounting of 1.2 million vehicles per year as a “major discrepancy”, the government remains a closed book on the subject.

After multiple emails back and forth, Webb says he’s not hopeful of getting a straight answer from TD. “[T]hey simply would not admit that there was any deficiency in the vehicle-counting system,” he says.