OBSERVED · MATCHED SEASONS
53/60
Longer mean journeys
Spring 2023 to spring 2024. The unweighted median change is +37.5 seconds; it is not the average delay per Welsh traveller.
Retrieving and verifying the requested analytical release. Result data loads before optional map boundaries and source checks.
Your filters and release stay in the address. Navigation remains available.
← STATS 19 research overview / Phase 4 group patterns / Phase 3 comparisons / Phase 2 observations / Phase 1 protocol
20MPH EVIDENCE LAB · PHASE 5 · 4 NOVEMBER 2026
Examine measured journey changes, then explore the casualty prevention needed to offset an assumed delay. The evidence does not yet support a national net-benefit figure.
Trade-offs · observed journeys and conditional scenarios · causal gate not met
EVIDENCE BEFORE A PRICE TAG
Certain: the published route observations and the valuation arithmetic can be reproduced. Unresolved: the policy’s share of the delay, a credible severity-specific casualty effect, and matching traffic volumes. The calculator exposes those assumptions instead of filling the gaps with a national estimate.
OBSERVED · MATCHED SEASONS
Spring 2023 to spring 2024. The unweighted median change is +37.5 seconds; it is not the average delay per Welsh traveller.
OBSERVED · SAME COMBINATIONS
The 95th–5th percentile spread became smaller. A longer journey can also be more predictable. Neither change is isolated as a policy effect.
SCENARIO · YOUR ASSUMPTIONS
Change the assumed traffic, delay and costs. Any casualty prevention you enter stays explicitly hypothetical.
MEASURED SPEEDS · 38 RETAINED SITES
The weighted mean fell from 28.5mph in summer 2023 to 25.2mph in summer 2025. Five reverted sites are excluded throughout; these through roads do not represent all Welsh roads.
Explore speed measures and all observation periods → · Exact speed data
T F W · 15 THROUGH ROUTES · 60 COMBINATIONS
Each combination holds route, direction and peak period fixed. Spring uses Tuesday–Thursday observations in February–April of 2023 and 2024, avoiding Easter holidays. Autumn uses September–November 2024 and changes the season. The latest journey release located at the 4 November 2026 check remains 15 July 2025.
53 longer, 1 unchanged, 6 shorter. Median mean-duration change +37.5 seconds. 50 narrower and 10 wider percentile spreads; 44 combinations became both longer and narrower. Every combination has equal weight; these are not traveller-weighted or independent observations.
Mean: circle. Observed 5th–95th percentile range: bar. This range describes journey variation; it is not a confidence interval, a standard deviation or an estimate of reliability benefits.
| Period | Mean | 5th percentile | 95th percentile | Spread |
|---|---|---|---|---|
| Spring 2023 | 64m 35s | 60m 31s | 69m 42s | 9m 11s |
| Spring 2024 | 67m 33s | 63m 49s | 71m 35s | 7m 46s |
| Autumn 2024 | 68m 23s | 64m 13s | 73m 15s | 9m 2s |
Selected same-season comparison: mean +178 seconds (+4.6%), spread -85 seconds. The source direction codes are retained without inventing compass directions.
| Route / peak / direction | 2023 spring mean | Selected later mean | Mean change | Spread change |
|---|---|---|---|---|
| Aberystwyth to Cardigan · AM · 1 | 3875 | 4053 | +178 | -85 |
| Aberystwyth to Cardigan · PM · 1 | 3870 | 4015 | +145 | +8 |
| Llandudno to Betws-y-coed · AM · 1 | 1768 | 1866 | +98 | -33 |
| Llandudno to Betws-y-coed · PM · 1 | 1791 | 1872 | +81 | -13 |
| Carmarthen to Coleg Sir Gar · AM · 1 | 1735 | 1836 | +101 | +35 |
| Carmarthen to Coleg Sir Gar · PM · 1 | 1728 | 1865 | +137 | +44 |
| Porthmadog to Caernarfon · AM · 1 | 1772 | 1821 | +49 | -32 |
| Porthmadog to Caernarfon · PM · 1 | 1806 | 1827 | +21 | -207 |
| Caerphilly to Newport · AM · 1 | 1627 | 1761 | +134 | -103 |
| Caerphilly to Newport · PM · 1 | 1610 | 1705 | +95 | -130 |
| Wrexham to Deeside (Shotton) · AM · 1 | 1630 | 1683 | +53 | -275 |
| Wrexham to Deeside (Shotton) · PM · 1 | 1658 | 1719 | +61 | -149 |
| St Davids to Haverfordwest · AM · 1 | 1642 | 1661 | +19 | +1 |
| St Davids to Haverfordwest · PM · 1 | 1661 | 1661 | 0 | +6 |
| Ystrad Mynach to Crosskeys · AM · 1 | 1027 | 1112 | +85 | -55 |
| Ystrad Mynach to Crosskeys · PM · 1 | 1032 | 1113 | +81 | -13 |
| Sketty to Morriston · AM · 1 | 1021 | 1024 | +3 | -156 |
| Sketty to Morriston · PM · 1 | 1099 | 1062 | -37 | -157 |
| Pontypool to Blaenavon · AM · 1 | 840 | 909 | +69 | +50 |
| Pontypool to Blaenavon · PM · 1 | 897 | 959 | +62 | -11 |
| Bangor-on-Dee to Wrexham Maelor Hospital · AM · 1 | 854 | 865 | +11 | -121 |
| Bangor-on-Dee to Wrexham Maelor Hospital · PM · 1 | 871 | 882 | +11 | -78 |
| Barry Hospital to Dinas Powys · AM · 1 | 858 | 857 | -1 | -153 |
| Barry Hospital to Dinas Powys · PM · 1 | 797 | 828 | +31 | -53 |
| St Asaph to Denbigh · AM · 1 | 753 | 790 | +37 | -45 |
| St Asaph to Denbigh · PM · 1 | 750 | 781 | +31 | -31 |
| Treforest to Nantgarw · AM · 1 | 724 | 779 | +55 | +14 |
| Treforest to Nantgarw · PM · 1 | 827 | 778 | -49 | -157 |
| Coedpoeth to Wrexham Maelor Hospital · AM · 1 | 500 | 532 | +32 | -33 |
| Coedpoeth to Wrexham Maelor Hospital · PM · 1 | 470 | 505 | +35 | -16 |
| Aberystwyth to Cardigan · AM · 2 | 3864 | 4010 | +146 | -52 |
| Aberystwyth to Cardigan · PM · 2 | 3853 | 4033 | +180 | -50 |
| Llandudno to Betws-y-coed · AM · 2 | 1914 | 1936 | +22 | -154 |
| Llandudno to Betws-y-coed · PM · 2 | 1912 | 1960 | +48 | -82 |
| Carmarthen to Coleg Sir Gar · AM · 2 | 1735 | 1899 | +164 | +97 |
| Carmarthen to Coleg Sir Gar · PM · 2 | 1732 | 1857 | +125 | +78 |
| Porthmadog to Caernarfon · AM · 2 | 1798 | 1886 | +88 | -9 |
| Porthmadog to Caernarfon · PM · 2 | 1814 | 1851 | +37 | -138 |
| Caerphilly to Newport · AM · 2 | 1536 | 1665 | +129 | -6 |
| Caerphilly to Newport · PM · 2 | 1738 | 1758 | +20 | -303 |
| Wrexham to Deeside (Shotton) · AM · 2 | 1722 | 1741 | +19 | -336 |
| Wrexham to Deeside (Shotton) · PM · 2 | 1708 | 1733 | +25 | -197 |
| St Davids to Haverfordwest · AM · 2 | 1653 | 1694 | +41 | -14 |
| St Davids to Haverfordwest · PM · 2 | 1654 | 1656 | +2 | -34 |
| Ystrad Mynach to Crosskeys · AM · 2 | 1090 | 1153 | +63 | -96 |
| Ystrad Mynach to Crosskeys · PM · 2 | 1107 | 1172 | +65 | -55 |
| Sketty to Morriston · AM · 2 | 1103 | 1055 | -48 | -113 |
| Sketty to Morriston · PM · 2 | 1110 | 1076 | -34 | -112 |
| Pontypool to Blaenavon · AM · 2 | 869 | 960 | +91 | -24 |
| Pontypool to Blaenavon · PM · 2 | 768 | 842 | +74 | +42 |
| Bangor-on-Dee to Wrexham Maelor Hospital · AM · 2 | 917 | 924 | +7 | -78 |
| Bangor-on-Dee to Wrexham Maelor Hospital · PM · 2 | 944 | 951 | +7 | -94 |
| Barry Hospital to Dinas Powys · AM · 2 | 783 | 808 | +25 | -72 |
| Barry Hospital to Dinas Powys · PM · 2 | 835 | 864 | +29 | -64 |
| St Asaph to Denbigh · AM · 2 | 782 | 815 | +33 | -58 |
| St Asaph to Denbigh · PM · 2 | 768 | 809 | +41 | -27 |
| Treforest to Nantgarw · AM · 2 | 601 | 639 | +38 | -26 |
| Treforest to Nantgarw · PM · 2 | 672 | 666 | -6 | -100 |
| Coedpoeth to Wrexham Maelor Hospital · AM · 2 | 473 | 506 | +33 | -29 |
| Coedpoeth to Wrexham Maelor Hospital · PM · 2 | 526 | 547 | +21 | -65 |
These selected through routes span multiple limits. No matched route traffic weights, person occupancy, journey-purpose mix or route-specific roadworks history has been validated. The report’s roadworks-free statement concerns speed-monitoring sites, not these journey routes. Do not multiply the sample median by national trip totals.
Official journey report · Original TfW workbook · All derived journey values
CONDITIONAL SOCIAL VALUE · NOT CASH SAVINGS
Travel time and casualty prevention are valued as social welfare. They are not forecasts of NHS spending, insurance claims or premiums. The threshold prices only the components listed here.
Illustrative car-travel cohort. Every volume, delay, occupancy and casualty input below is an assumption. No observed journey change or phase 3 association is automatically treated as a policy effect. This does not estimate the balance for Wales.
HYPOTHETICAL SCENARIO · 10 YEARS · £2023 PRICES
Hypothetical cohort, not a measured Welsh policy cost. These results use your inputs. No national net benefit is established.
25,000,000 vehicle trips and 312,500 additional person-hours each year. Time value: £44.39m; upfront cost: £0; recurring cost: £0. Present values are discounted to 2026.
In each year, an alternative equivalent of 1.78 fatal, 16.03 serious or 208.8 slight casualties prevented. These are three separate ways to reach the threshold, not quantities to add together.
| Assumed seconds per trip | Included cost present value | Serious casualties per year needed to offset it |
|---|---|---|
| 0 | £0 | 0.00 |
| 15 | £22.20m | 8.01 |
| 30 | £44.39m | 16.03 |
| 60 | £88.79m | 32.05 |
Use net changes by severity for the same cohort and years. Negative prevention means additional casualties. Serious and slight changes cannot be inferred from the lab’s unadjusted recording series. No reduction is selected by default.
No casualty effect entered. A policy balance is withheld.
Includes the exact assumptions, annual discounted components, valuation source and limitations. These results remain labelled as hypothetical when exported.
{
"assumptions": {
"tripsPerDay": 100000,
"daysPerYear": 250,
"occupancy": 1.5,
"delaySeconds": 30,
"purpose": "commute",
"years": 10,
"upfrontCost": 0,
"annualCost": 0,
"fatal": 0,
"serious": 0,
"slight": 0,
"benefitStatus": "Not entered; zero placeholders do not assert no policy benefit"
},
"result": {
"annualTrips": 25000000,
"personHours": 312500,
"timePV": 44393556.862851575,
"operatingPV": 0,
"upfrontPV": 0,
"costPV": 44393556.862851575,
"safetyPV": null,
"componentBalancePV": null,
"breakEven": {
"fatal": 1.776561426632536,
"serious": 16.02548606970332,
"slight": 208.84737473699255
},
"casualtyPV": {
"fatal": 24988472.786443055,
"serious": 2770184.733852097,
"slight": 212564.59133737086
},
"annual": [
{
"year": 2027,
"timePV": 4839187.447168911,
"operatingPV": 0,
"safetyPV": null,
"componentBalancePV": null
},
{
"year": 2028,
"timePV": 4745676.578624583,
"operatingPV": 0,
"safetyPV": null,
"componentBalancePV": null
},
{
"year": 2029,
"timePV": 4653972.683385461,
"operatingPV": 0,
"safetyPV": null,
"componentBalancePV": null
},
{
"year": 2030,
"timePV": 4564040.844092988,
"operatingPV": 0,
"safetyPV": null,
"componentBalancePV": null
},
{
"year": 2031,
"timePV": 4475846.818120177,
"operatingPV": 0,
"safetyPV": null,
"componentBalancePV": null
},
{
"year": 2032,
"timePV": 4389357.024533313,
"operatingPV": 0,
"safetyPV": null,
"componentBalancePV": null
},
{
"year": 2033,
"timePV": 4304538.531305617,
"operatingPV": 0,
"safetyPV": null,
"componentBalancePV": null
},
{
"year": 2034,
"timePV": 4221359.042777972,
"operatingPV": 0,
"safetyPV": null,
"componentBalancePV": null
},
{
"year": 2035,
"timePV": 4139786.887361972,
"operatingPV": 0,
"safetyPV": null,
"componentBalancePV": null
},
{
"year": 2036,
"timePV": 4059791.005480582,
"operatingPV": 0,
"safetyPV": null,
"componentBalancePV": null
}
]
}
}MEASURED, REPORTED AND MISSING
The dated figures below have different scope and status. They are neither additive nor converted into the calculator’s price basis.
| Source / date | Reported amount | What it supports |
|---|---|---|
| 2022 regulatory forecast 2022-06-21 | £32.500m | Forecast, not expenditure. Table 4 note: discounted 2022-price present value over a 30-year appraisal. Header says current prices; preserve that ambiguity. Main expected spending in 2022–2027. Excludes already-sunk taskforce and RIA costs. |
| February 2025 disclosure 2025-02-10 | £34.692m | Mixed paid and awarded. Earlier financial years claimed and paid; FY2024/25 awarded. No constant price basis stated. Pilots, local authorities, communication, consultancy, research and marketing. Not a final audited national outturn. |
| May 2025 spending disclosure 2025-05-19 | Around £39.504m | Approximately spent to date; provisional. FY2024/25 outturn still being finalised. No common price year or annual payment schedule supplied. Includes the separate FY2024/25 Operation Ugain grant. Earlier 20mph policing cannot be separated from normal operations. Do not add the grant again. |
| Current government FAQ 2026-03-03 | Around £32.000m | Rounded implementation figure. No accounting period or price year given. The FAQ remains around £32m; its scope is not reconciled with the broader May 2025 disclosure. |
| Funding for revised limits 2024-07-16 | £5.000m | FY2024/25 funding availability. Allocation announcement, not actual expenditure. Support for authority changes under revised guidance. Do not add to cumulative spending without reconciling overlap. |
| Published FY2025/26 awards 2025-04-11 | £5.107m | Sum of 17 named grant awards. Awards, not measured expenditure. Sum is derived from the 17 explicitly labelled 20mph implementation rows. Missing authority entries are not zero; not necessarily a complete national programme cost. Do not add directly to cumulative spent-to-date figures. |
For a future retain-or-reverse decision, already-spent implementation costs are sunk; an appraisal of the historical policy needs costs on their actual dates. The phase 3 identification checks remain unresolved. The lab does not convert the aggregate casualty association into avoided deaths or serious injuries, or multiply it by a national average casualty value.
OPTIONAL EXTENSION · COVERAGE GATE NOT MET
The existing TfW pilot did not collect a pre-implementation air-quality baseline. Sensor relocation, construction, calibration gaps and incomplete comparison-site traffic information limit its post-only comparisons. That does not establish either a policy effect or no effect.
Hourly pollution and quality flags are publicly available, but dated treatment/control pairs, usable before/after capture and suitable local weather have not been verified for a national study. The optional causal analysis therefore does not proceed.
TfW pilot air-quality report. NO₂ and particulates are separate outcomes from carbon emissions; ambient concentrations do not directly measure fuel use or CO₂.
REPRODUCIBLE RELEASE · V5
The 60 journey combinations reconcile to the frozen phase 2 extract and original workbook checks. The calculator uses the May 2026 DfT TAG Data Book, with 2023 market prices and valuations uprated to 2026. The chosen present-value reference is 2026; future annual flows occur at year end from 2027. This is a transparent illustration using TAG values and conventions, not a complete TAG appraisal.
| Unit | Value |
|---|---|
| Car business travel · per person-hour | £28.06 |
| Commuting · per person-hour | £15.79 |
| Other personal travel · per person-hour | £7.21 |
| One fatal casualty prevented | £2,713,825 Human: £2,541,353; output and medical: £172,472 |
| One serious casualty prevented | £301,677 Human: £248,781; output and medical: £52,896 |
| One slight casualty prevented | £23,171 Human: £18,211; output and medical: £4,960 |
Human casualty values remain fixed in real terms and are discounted at 1.5%. Output and medical values, and travel-time values, grow by 1.5% a year and are discounted at 3.5%. Entered recurring resource costs stay constant in real terms and use 3.5%. Horizons stop at 30 years; later discount-rate changes are outside this calculator.
Annual person-hours = vehicle trips per day × days per year × people per car × assumed seconds per trip ÷ 3,600. The safety threshold divides included cost present value by the present value of preventing one casualty of the chosen severity in every year. No averaging of severity values or adding collision totals is used.
The 2026 income index uses a forecast. Car business travel uses an all-distance average; a full appraisal may need distance-dependent values. No reliability, air pollution, carbon, walking/cycling, vehicle operating, damage-only, police, insurance administration or wider economic effect is valued. Missing components are not zero effects. Medical costs already sit within casualty values: adding a separate NHS saving would double count.
Upfront costs occur in 2026; no residual asset value is assumed. Monetary inputs must be in the same market-price unit, with any necessary inflation and indirect-tax conversion performed first. The DfT’s usual reporting present-value base is 2023; our clearly labelled 2026 reference is a presentation choice.
DfT TAG Data Book and release history · Exact cells, components and formulas · Source register and retrieval audit