When scouring the Internet for Hong Kong news, some of my favourite stories are those highlighting the latest research findings from their highly ranked universities. During the pandemic I posted a lot of Covid-19 research, so nowadays it's refreshing to see studies on other topics whose findings are accessible to a general audience:

The Hong Kong government, the masters of bureaucratese, call them "auxiliary devices for pedestrian safety." When the light is red, the devices project red lights from above towards the ground, with the hope that a pedestrian looking down at their mobile phone will see the red light even when distracted enough not to be able to see the traffic signal ahead with the standing red man.

The link to the article is here, but as it is a paper written for an academic journal, it is paywalled, and if you know anything about the "academic biz," then you'll know that journals charge an arm and a leg for access to articles. Perhaps someone from HKU who is reading this could help out! Anyway, it doesn't matter, as the abstract of the paper tells plenty.

This led to various snarky replies on both Bluesky and the HellSite™, ranging from Hong Kong is becoming a red light district, to those devices are ugly, to the red light is ugly, to this:

The study's findings were based on the 6-month trial beginning in July 2022 in which red-light projection devices were installed at four intersections. Afterwards, the Transport Department proceeded to install red-light projectors in 100 locations throughout Hong Kong starting in November 2023 and ending in December 2024. A team from HKU analysed 20 of those locations, and they yielded virtually the same result: the proportion of pedestrians running red lights at those intersections fell 24% after the devices were installed.

The same LegCo Transport paper provides data on the number of pedestrian accidents at those 100 locations before and after the red-light projection devices were installed.

A 73% decrease in the number of average pedestrian incidents per location in 2025 compared to the years 2020 through 2023 is certainly an impressive result, with the accident rate per targeted location nearly identical to the other 1,900 locations without red-light projectors. But pedestrian accident rates fell 37% in those other 1,900 locations at that time period, suggesting that the devices are not solely responsible for the 73% decrease.

In any case, red-light devices seem to reduce pedestrian accidents at traffic junctions, but are they cost-effective? According to papers from Hong Kong's Legislative Council, the installation of red-light devices at the 100 locations cost HK$21.66 million, plus HK$330,000 for the initial HKU study. (The cost of the subsequent HKU study has not been disclosed.) Tack on annual costs of about HK$260,000 in electricity, and HK$100,000 in maintenance, and the total cost over, say, a ten-year lifespan, including inflation, will be over HK$30 million.

Unfortunately, there isn't enough data to determine whether the benefits outweigh the costs; heck, there isn't even a a standardised, Hong Kong-specific framework for assigning economic values to fatalities and injuries in road accidents that I am aware. I spent half an hour trying to estimate the benefits and my terrible calculations suggest they outweigh the costs; in which case, couldn't they have built a red light device that isn't fugly?

(Then again, my confidence in their ability to build eye-pleasing light structures is low at this point, considering what the government tried to pull off recently.)

Look in the mirror, drivers

On a slightly separate topic, a Numberwang reader pointed out a government press release which was part of questions from LegCo to the government: Duncan Chiu asked whether penalties for pedestrians running red lights (up to HK$2,000) and littering (a fixed penalty of HK$3,000) are too inflexible. As part of the answer, the Secretary for Transport and Logistics Mable Chan provided figures of pedestrian casualties in traffic accidents in the past three years. The numbers are interesting!

The first big observation is the contributory factor to accidents in which pedestrians "cross the road heedless of traffic at crossing" in the past three years: they account for just 3.87% of all pedestrian casualties; 7.05% of all pedestrian deaths, 5.62% of all serious injuries, and 3.61% of all minor injuries.

The other big observation is that from 2023 to 2025, 77% of all pedestrian casualties occurred in which there was "no pedestrian factor," including two out of three pedestrian deaths. Even after the 100 red-light devices were installed, pedestrian deaths with "no pedestrian factor" across Hong Kong jumped 48% in 2025.

Does "no pedestrian factor" mean "it's the driver's fault"? There are other reasons that might fall outside the category, but I would speculate nearly all "no pedestrian factor" casualties are driver-related, and if the Transport Department is serious about reducing pedestrian casualties, perhaps they need to concentrate on the "no pedestrian factor" a little bit more with some sensible traffic-calming solutions.

🙏 My thanks to James Ockenden for contributing.