Japan Cut the Default Speed Limit on Residential Roads to 30 km/h — a Nationwide Noise Measure Enacted Without Noise as an Objective

This note is the institutional analysis part of the Quiet Neighbourhood Project. On the fragmentation of regulation see the structural analysis of regulatory gaps（このサイトの記事）; on the level of the standards themselves see the WHO 2018 environmental noise guidelines against Japan's standards（このサイトの記事）.

On 1 September 2026 the default speed limit on residential roads was lowered from 60 km/h to 30 km/h.

The scope covers ordinary roads with neither a centre line nor marked traffic lanes, described as roads used mainly for the daily life of local residents[1]. Excluded are ordinary roads with a centre line or traffic lanes marked by road signs or road markings[1] and ordinary roads where opposing traffic is separated by road structure, fences or other installations[1].

What distinguishes this change is that it does not depend on posting signs. Where the shape of the road meets the condition, 30 km/h is the legal limit whether or not anything is posted. It took effect on every qualifying road in the country on the same day.

The purpose is stated as well. The Tokyo Metropolitan Police give it as securing traffic safety and smooth flow, and protecting drivers, passengers, pedestrians and cyclists[1].

Noise appears in neither the National Police Agency's nor the Tokyo Metropolitan Police's explanation.

Speed and sound are joined by an equation

Road traffic noise in Japan is predicted using the model compiled by the Research Committee on Road Traffic Noise. The basic equation appears in a reference document of the National Institute for Land and Infrastructure Management.

The A-weighted sound power level of vehicle running noise, LWA [dB], is calculated by the following equation: LWA = a + b log10 V + C[2]. V is running speed, a is a constant given by vehicle class, and b expresses the dependence on speed.

The value of b is fixed. The coefficient b expressing speed dependence is 30 for steady-running sections and 10 for non-steady-running sections, for every vehicle class[2].

The shape of the equation says something simple. Sound moves with the logarithm of speed. Lower the speed and the sound falls.

Which of the two sections a residential road belongs to has to be checked, though. The model defines them like this. Steady-running section: a motorway, or an ordinary road sufficiently distant from a signalised intersection, where vehicles run in or near top gear. Running speed V is taken over the range 40 km/h to 140 km/h[2]. The other is Non-steady-running section: an ordinary road containing signalised intersections, where vehicles repeatedly accelerate and decelerate. Running speed V is taken over the range 10 km/h to 60 km/h[2].

Thirty km/h falls below 40 km/h, the lower bound of the steady-running section. The speeds expected on residential roads therefore sit on the non-steady side, where the coefficient is 10 rather than 30. The drop in sound for a given drop in speed is smaller than the coefficient of 30 would imply.

That is consistent with the European measurements considered next. Where cities cut limits from 50 km/h to 30 km/h, the reductions reported mostly fall within a few decibels.

Only one purpose is written down

The leaflet the National Police Agency distributed to drivers says the same and no more. Speed limits exist to secure traffic safety and smooth flow, and observing them helps protect drivers, passengers, pedestrians and cyclists[3].

This is a change made for safety, and there is nothing to dispute in that.

The published explanation stops there. No figure showing why the number is 30 km/h appears on the police pages or in the leaflet. The shape of the roads covered, the speed to observe, and a single line of purpose. That is all.

What follows from having only one purpose is the issue. One purpose produces one way of measuring. Collision counts and casualty numbers are compiled every year. Nothing in this change compiles what happened to the sound on residential roads.

Other cities used the same instrument for noise

Using speed limits as a noise measure is not unusual outside Japan.

A review by Michelaraki and Yannis of the National Technical University of Athens gathers results from 40 European cities. Zurich appears among them. Zurich's decision to lower its speed limit was primarily driven by the goal of reducing noise pollution. After implementing this measure, traffic noise decreased by an average of 1.6 dB during the day and 1.7 dB at night[4].

Zurich lowered speeds in order to reduce noise, and measured falls of 1.6 dB by day and 1.7 dB at night. The same review lists other cities. in Buxtehude, Germany, there was a reduction of 7 dB. Similarly, in Graz, Austria, noise levels decreased by 1-2 dB, while in Berlin, Germany, a reduction of 3 dB was observed. In Modena, Italy, noise levels decreased by 3-5 dB[4].

On scope and speed of application, Japan's approach is broader and faster. While Zurich picks off streets that exceed the limit one at a time, Japan switched every qualifying road in a single day. In sheer scale of the instrument, this is likely among the largest such changes anywhere.

What is not being measured is the sound.

What changed is the ceiling, not the actual speed

Some care is needed here. Lowering the legal limit from 60 km/h to 30 km/h is not the same as cars actually slowing down.

Nobody was driving residential roads at close to 60 km/h as a rule. On narrow roads with poor sightlines and people walking, speeds already sat below the limit. So how much actual speed falls is not settled in advance. It depends on how far the limit is observed.

From the acoustic side this is decisive. What enters the equation is running speed V, not the legal limit. Rewriting the ceiling does not by itself move V.

Which is exactly why it has to be measured. How much did speeds fall, and how did that fall show up in the sound? Both remain estimates until someone measures.

Zurich's 1.6 dB is itself a measured value rather than a model output. It is a field result that already contains whatever speed reduction actually occurred.

What is not measured cannot be counted as an effect

As the complaint gap（このサイトの記事） set out, sound on residential roads rarely leaves a record in the first place. If reporting changes nothing, not reporting is the rational choice, and complaint counts stop representing the amount of harm.

The same thing can happen here. If the sound falls, nothing is created anywhere to show that it fell. If it does not fall, that too goes unrecorded. Either way, this change will be judged solely on whether collisions decreased.

As the environmental justice hypothesis for roadside residents（このサイトの記事） laid out, Japan's environmental standards set looser values for areas facing roads and allow long grace periods for meeting them. Residential roads tend to fall outside continuous monitoring altogether. A change that happens where no measurement point exists does not appear in the statistics of the system.

What having one purpose costs

The objection here is not to putting safety first. The evidence that pedestrian fatality rates change around 30 km/h is clear, and that reason alone justifies the change.

The point is different. A secondary benefit that is not written in as an objective goes unmeasured, and therefore cannot inform the next decision.

Three things are lost in concrete terms.

First, only one line of evidence remains for defending the rule. Sooner or later the limit will be judged on whether collisions fell. Collisions are rare events with large year-to-year variation, so the effect is hard to see over a short window. Sound is a continuous quantity and shows differences over shorter periods. An indicator that could have carried the argument has been set aside in advance.

Second, residents lose the words for what changed. A sense that the street got quieter becomes a request to the authorities only when it attaches to a measurement. Without numbers it stays a private impression.

Deciding what counts as a result before the work starts is the subject of designing outcome indicators（このサイトの記事）. An indicator can be added later; the value from before the change cannot.

Third, less is carried forward into the next design. How speed and sound actually relate on Japanese residential roads can only be learned from a change of this kind. Miss this one and there is no telling when a natural experiment of the same scale comes round again.

What this change holds as an occasion for observation

The change took effect on 1 September 2026, so the chance to measure the before state has passed.

Things remain observable nonetheless. Existing continuous monitoring points near residential roads hold time series that run unbroken across the change. A difference is already recorded there, and it can be read back later.

The other is that compliance shifts over time. Observance of a speed limit is typically highest just after it takes effect and loosens afterwards. If the sound rises again as compliance loosens, the speed–sound relation was holding on Japanese residential roads too. If it does not move, that indicates actual running speeds were never bound by the ceiling in the first place.

So this change has created, at nationwide scale, a state in which the relation between speed and sound can be observed. Among the hypotheses this lab has been working with, the complaint gap and the relationship between land-use zoning and noise inequality set out in the four research hypotheses（このサイトの記事） have both lacked measurements on residential roads. Against that gap, this amounts to a natural experiment with the conditions in place.

An instrument that lowers sound was deployed without naming sound.

Who is going to check whether it worked?

References

The default speed limit for vehicles on residential roads is being lowered — National Police Agency (2026). National Police Agency

On the Default Speed Limit for Residential Roads — Tokyo Metropolitan Police Department (2026). Tokyo Metropolitan Police Department

Road Traffic Noise Prediction Model ASJ RTN-Model 2013 — Research Committee on Road Traffic Noise, Acoustical Society of Japan (2013). NILIM reference document

Assessing the effectiveness of 30km/h speed limit in cities — Michelaraki, E. and Yannis, G. (2024). National Technical University of Athens