humancentric

Room enough to walk two abreast.

Beside a cycling path the standard is 1.5 metres. Two adults are 1.22 metres of shoulder before anyone has made room for anything. This is the case for closing that gap, argued from Singapore's own documents.

This is a 1.5 metre footpath.

A drawing, not a render. Drag to walk around it. The bar across each figure is a 0.45 m shoulder line and the ring on the ground is the space that person actually occupies. Nothing here is scaled to look convincing.

Watch the rings. They meet long before the shoulders do, and that is what a cramped footpath actually feels like — not collision, but the constant low-grade negotiation of space with whoever you are walking with.

Space was found for cycling paths. Pedestrians were made to contribute anyway.

The LTA Walking and Cycling Design Guide sets a standalone footpath at 1.8 m. Section 2.1.1 explains why, in LTA's own words:

"With Singapore's ageing population, the use of Personal Mobility Aids (PMAs) is becoming more prevalent. New footpaths have been increased from 1.5m to 1.8m, which is the adequate width for two typical wheelchairs to pass each another [sic]."

§2.1.1

Section 2.1.2 then withholds that uplift wherever a cycling path is present:

"Cycling paths adjacent to footpaths may be used to negotiate manoeuvring space required by pedestrians, including wheelchair users. Under such circumstances, 1.5m is adopted as the dedicated pedestrian width."

§2.1.2

The part that makes this hard to defend

It would be one thing if there were simply no room. But the guide's own typology drawings show that room was found.

6.8 mroadside table without a cycling path: 1.8 m footpath + 2 m planting verge + 3 m service verge§3.1.1
8.5 mwith one: 1.5 m footpath + 2 m cycling path + 2 m planting + 3 m service§3.1.2
−0.3 mchange to the footpath, while the table around it grew by 1.7 mDerived

The roadside table widens by 1.7 m to accommodate a 2 m cycling path. So 1.7 of the 2 metres came from outside the pedestrian envelope. Space was found. Land was committed. And the footpath was still cut by 300 mm.

That is the question this puts. Not "why is Singapore so cramped", which has an obvious and sufficient answer about land. The question is narrower and harder: when 1.7 metres was already being found elsewhere, why did pedestrians have to contribute the last 300 millimetres at all?

It applies it on bridges, and not on the ground.

"The typical clear width of the footpath on the structure is 2m and it is determined based on a space for two pedestrians to walk side by side and a third pedestrian to overtake."

Same document, §5.2.1, on pedestrian overhead bridges

LTA has done the arithmetic for sociable walking and written down the answer. Note also that the bridge figure is a clear width, 2 m of walking space inside a 2.2 m structure, whereas the at-grade 1.5 m and 1.8 m figures are nominal paved widths that include lamp posts, bins and drain covers. Singapore applies its most generous and most carefully defined pedestrian width when people are on a bridge, and its least generous when they are on the ground.

A gap worth naming. BCA's Code on Accessibility publishes a table of comfortable access widths: two wheelchairs passing, 1800 mm; a wheelchair and a person, 1500 mm; a blind person with a sighted guide, 1200 mm. There is no entry anywhere in the Code for two ambulant adults walking side by side. Singapore's standards recognise accompanied walking only when one of the two people is disabled.

Most people are not walking alone, and they walk abreast in order to talk.

Moussaïd and colleagues, in PLoS ONE (2010), tracked 1,353 pedestrian groups at two sites: a public square in Toulouse on a weekday afternoon, and a commercial walkway on a Saturday.

55–70%of pedestrians were in a group, not alone — 55% weekday, 70% SaturdayObserved
0.78 mmeasured centre-to-centre spacing of a walking pair at low densityObserved
17%walking-speed reduction the authors' calibrated model predicts when the preference to communicate is switched onSimulated

The paper is explicit about why people arrange themselves this way: the side-by-side formation "facilitates social interactions within the groups because each member can easily communicate with his partners without turning the back to any of them."

When space tightens, the formation deforms. Zanlungo et al. (Physical Review E, 2015) found group extent perpendicular to travel falls linearly as density rises. In a controlled corridor experiment, Hu et al. (Safety Science, 2021) found groups reorganised into single file approaching a narrow exit. Jan Gehl described the same effect four decades earlier: "People no longer meet but walk behind one another in ranks."

Where the evidence stops, and this page says so. No published study measures real footpath width against the observed rate at which people walk in pairs or hold a conversation. The studies above measure the effect of density, and in Hu's case a bottleneck. The chain here is a strong inference across three bodies of evidence, not a measured finding, and nothing here claims otherwise. The width case does not depend on it: the sections above stand on anthropometry and on LTA's own accessibility reasoning alone.

At the bottom of the range, and worse than the numbers suggest.

City / countryMinimumMeasured asInstrument
Singapore — beside a cycling path1.5 mNominal pavedLTA Design Guide §2.1.2
Taiwan — constrained, with approval0.9 mClear (淨寬)Urban Roads Design Standards, Art. 16(1)
Taiwan — roads under 12 m wide1.2 mClearArt. 16(1)
Taiwan — standard minimum1.5 mClearArt. 16(1)
South Korea — topographic exception1.5 mEffective (유효폭)Road Structure Rules, Art. 16
Singapore — standalone footpath1.8 mNominal paved§2.1.1
Japan — ordinary road2.0 mClear, plus mandatory additionsRoad Structure Ordinance, Art. 11(3)–(4)
South Korea — standard2.0 mEffectiveArt. 16
Netherlands — basic network2.0 mClear walking spaceCROW Ontwerpwijzer voetgangers, level B
Netherlands — main network2.9 mClear walking spaceCROW, level A
Hong Kong — urban3.0 mThrough-zone clearTPDM Vol 6 §10.5.5.2
Japan — heavy pedestrian traffic3.5 mClear, plus additionsArt. 11(3)

The "measured as" column is the point of the table. Every comparator specifies a clear or effective width, exclusive of street furniture. Singapore's are nominal paved widths that include lamp posts, bins and drain covers. Japan additionally requires width to be added for roadside features: +0.5 m generally, +1.0 m for a bench, +1.5 m for street trees, +3.0 m at an overbridge. Hong Kong's 3.0 m sits on top of a 0.5–1.0 m frontage zone and a 2.0–3.0 m furniture and greening zone. The gap runs against Singapore in every row.

Choose the comparison class deliberately. Singapore benchmarks itself against Hong Kong, Taipei, Seoul and Tokyo on rail reliability and mode share. On street design its own documents reach elsewhere: the Walking and Cycling Design Guide's foreign references are AASHTO, Sustrans, Eurocode and a UK bridge-design text, with no foreign Asian standard among them, and the LTMP 2040 Advisory Panel report contains exactly one international case study, "Healthy Streets: Lessons from London." Comparing footpath widths to Asian peers is a departure from Singapore's own practice in those two documents. A fair departure, but it should be made knowingly.

Four changes, each specific enough to cost and to audit.

1 · Restore 1.8 m where a cycling path is present

Amend §2.1.2 so the dedicated pedestrian width does not fall below the 1.8 m set in §2.1.1. The §3.1.2 typology already widens the roadside table by 1.7 m to fit a 2 m cycling path. Take the last 300 mm from the same place as the other 1,700.

2 · Adopt 2.0 m clear as the target on routes with significant pedestrian volume

Use LTA's own §5.2.1 figure, and use it as LTA defines it there, as a clear width. It is already Singapore's number. It needs to mean the same thing at ground level that it means on a bridge.

3 · Specify a forgiving edge between footpath and cycling path

A shallow sloped kerb: detectable underfoot and by cane, but traversable. The argument for this is on its own page.

4 · Publish effective width, not nominal width

Require widths to be stated after deducting kerb and building-line clearances and street furniture, as Transport for London does and as every standard in the table above already does. A nominal 1.5 m path bounded by a wall and a kerb delivers roughly 1.1 m. This costs nothing to implement and it makes every other number in the system honest.

Where these go. LTA owns 1, 2 and 4. Number 3 touches LTA and the town councils. The existing policy vehicle is Friendly Streets, which already runs resident consultation and is committed to all 24 towns by 2030. Asking it to add a width criterion is smaller than asking for a new programme.

One caveat to raise before someone else does: the Walking and Cycling Design Guide is Public Version 1.0, content updated November 2018. The Active Mobility Design Guide v1.1, updated April 2023, still specifies a 1.5 m footpath alongside a 2 m cycling path, so the standard does appear current. Confirm that before submitting anything.

The four strongest objections, stated fairly.

Land is genuinely scarce. Every metre of footpath is a metre not available for housing.

True, and it should be conceded without hedging. Singapore has 744 km² for roughly six million people. The constraint is real, not a pretext.

But it does not answer the question, because nothing here asks for more land. It asks for a different split of land already committed to the road reserve, and §3.1.2 shows that when a cycling path was added, 1.7 m of additional width was committed. The land constraint did not prevent that. It only decided who paid the last 300 mm.

Singapore's footpaths are wide by Asian standards. You're comparing against the wrong cities.

Partly fair. Taiwan permits 0.9 m where constrained, so on paper there is a legal floor below Singapore's. Against Hong Kong's 3.0 m urban through-zone there is not.

The stronger point is not which city wins but what the numbers count. Every comparator specifies clear or effective width; Singapore specifies nominal paved width. A 1.5 m Taiwanese clear width is more walking space than a 1.5 m Singapore nominal width, and the table understates the gap for that reason.

Singaporeans socialise in malls, void decks and parks. The street isn't the social space here.

There is something to this and it deserves better than dismissal. Climate matters. Sheltered, air-conditioned space is more pleasant for a long conversation than a kerbside footpath at two in the afternoon.

Two responses. It partly begs the question: if the street is unpleasant to walk two abreast on, some of the retreat indoors is caused by the thing under discussion. And more usefully, the width case does not rest on sociability at all. It rests on anthropometry and on LTA's own accessibility reasoning. Sociability is why people notice the problem. Accessibility is why the standard is already hard to defend.

Widening every footpath in Singapore is prohibitively expensive.

Correct, and nobody is proposing it. Retrofitting the existing stock wholesale is not a serious ask.

What is asked is that the standard change, so new construction and scheduled renewal build to it. Standards work slowly and cheaply because they ride existing capital cycles. Change 3 costs close to nothing at resurfacing; change 4 costs nothing at all, being a reporting convention.

Sources

Singapore standards and policy

Anthropometry and pedestrian standards

  • Fruin, J.J. (1971), Pedestrian Planning and Design — the 0.61 × 0.46 m body ellipse
  • AASHTO, Guide for the Planning, Design and Operation of Pedestrian Facilities, 1st ed. (2004) §2.2.2 — "Two people walking side-by-side or passing one another generally require 1.4 m [4.67 ft] of space"
  • Transport for London, Pedestrian Comfort Guidance — 0.6 m body ellipse; 200 mm edge deductions
  • CROW, Ontwerpwijzer voetgangers — walking-width levels

Walking behaviour

Comparator standards

Known limits. Three things are marked as inference rather than finding and should stay marked. The link between footpath width and observed sociability is an inference across three bodies of evidence. The 1.1 m effective figure applies TfL's deduction method to a Singapore nominal width; it is a derivation, not a measured Singapore value, which is what the field protocol exists to fix. And the claim about Singapore's street-design documents looking to non-Asian sources is verified for the two documents named, not for LTA's entire corpus.

One measurement from you is worth more than another paragraph from us.

Nobody has ever measured Singapore footpath width against the rate at which people break formation. The protocol is written and waiting.