How to make safety decisions - Wales urban 20mph

[IvanV]

A bit of an essay here from me, on a bugbear of mine.

A few weeks ago I was circulated an article written by a road safety expert, where he objects to the presentation of "sensible speed limits" as "war on motorists".

In particular, we are told that road travellers in Wales have suffered "trifling inconveniences" for the "safety benefits" of the near-blanket urban 20mph speed limit. He makes the fair point that it is often Other People who suffer a large part of the safety risk - he mentions pedestrians, but could also mention People in Other Vehicles - which those presenting the risk so aren't sufficiently careful about. We don't really know what the safety benefits are, aside from a reduction in road speed, because it will probably take a few years to work out with acceptable statistical precision what has been the impact of the policy on road casualties. At a regional level, road casualties are sufficiently infrequent and clustered you can't really see where they are going on a month to month basis, or from one year to the next, it takes a few years.

But I was really interested in this issue of "trifling inconveniences". Are they really "trifling"? How big are they? And how might they compare with the safety benefits. Sadly, but unfortunately typically, the author seemed to have made no attempt at quantification. And the government does gives us values to do this quantification. You can disagree with those values, but they are what we are told to use to assess expenditure and policy and decide whether it is value for money. So it is a good starting point. When I read this article, I thought I'd do a back-of-the-envelope, but I couldn't locate data even to make an educated guess. It turns out that other people were a bit upset about this lack of data too, and made a noise about it, that got into the news. And so, just over a week ago, the Welsh government published its calculations on the subject. Which now gives me enough for my envelope reverse.

So we see that the total journey time extension the Welsh Government projected as a result of these lower road speeds is 1.0 bn person minutes, spread across for 1.2 bn person-journeys. That takes account of the fact that there may be multiple people in some vehicles.

This information enables us to value the journey time extension using the DfT's TAG (Transport Analysis Guidance) traveller values of time, which are in the TAG data book. But to do that we need to split travellers by journey purpose, because we are only given values of time by journey purpose, not an average. Unfortunately I couldn’t find average Welsh work/commute/other road journey purpose proportions, so I used some typical English proportions (non-London) for this, rounded off. Those came from the 2022 National Travel Survey for England. So I came up with a value, using 2023 values of time from TAG of £110m. Clearly this is a bit rough, but probably within 10%. Now a curiosity of TAG is that everything is valued in 2010 pounds. I could put an inflation rate on that to bring it to 2023, but I decided it was simpler just to do all my analysis in 2010 pounds.

Is £110m “trifling”? We can compare it to the social cost of Welsh road casualties. In 2022, there were 95 deaths, 921 serious injuries, 3431 other injuries in Wales. We can use TAG casualty values, and other accident social costs that it quantifies, to value the total cost of casualties - including other accident costs such as damage to vehicles and road furniture and police costs - on Welsh roads. That came to £400m, again denominated in 2010 money. That’s the total.

Many people don't like doing this social cost/social value type of argument where the values include the values of death and injury. But the point is, we only have limited resources in our society, and we need to decide where best to place them. Only by consistently analysing in this way can we address our resources to the places where they present highest value.

So the policy takes £110m in value from people, to try and reduce a social cost which is currently £400m.

What kind of reduction might we get? Road casualties in Wales would have to reduce by about 29% to justify the extra travel time (I used the unrounded figures for that ratio). That’s large, so I wouldn’t call that a “trivial” travel extension for amount of potential safety benefit. Urban roads in GB more extensively are responsible for around 65% of casualties, but only about 35% of fatalities, (that comes from some info reported by the RAC Foundation) which are a material part of the total cost. So on average casualties have a lower seriousness in urban areas, so let's guess that the amount of social cost of road casualties in urban areas is roughly half of that £400m. So that £110m is something around half the total urban casualty social value. (I'm focusing on urban because the 20mph only changes speeds in urban areas.) This measure would have to be quite extraordinarily successful to deliver anything like that. In all likelihood, the social cost of this policy is considerably higher than the social benefit.

But there is an important cost of accidents that is missing from TAG, and that is the cost of disruption to other travellers that occurs when there is an accident. The actions in clearing up an accident reduce road capacity, and there is queuing and congestion. When there is a bad accident on a motorway - not relevant in this case - that closes the road for several hours, the congestion can be monstrous, but this is relatively uncommon. This is not taken account of in the values for the social cost of accidents reported in TAG. Taking account of it would increase the social cost of accidents. TAG has no evidence on this. On a quick search, I couldn't find any studies addressing it, which is a significant lacuna in the research into social costs of travel. I have long tended to think that accident costs are underrated because of the failure to take this into account, especially on the railway where accidents are usually much more disruptive than on the roads. Potentially taking this into account might in principle make a large difference. It depends what the disruption costs are, and how much accident reduction results. But my suspicion would be that urban road accidents have relatively low disruption costs in comparison to, say, motorway accidents. So in fact the disruption avoided by a measure that focuses on minor urban roads might be small, and so the difference might not be large. But really this is something we need better evidence on.

A difficult argument. But to use “trifling” without any attempt at quantification is a logical fallacy, proof by assertion if you like. Such a word requires justification, and the road safety expert made no attempt to justify it. Unfortunately, failure to consider carefully the proportionality of measures to increase safety is a widespread failing in this country. We spend a disproportionate amount of money and time on relatively small gains in safety, while failing to address the difficult stuff. It means that our public expenditure is badly distorted. Other expenditure that could increase safety and people's health is reduced, while we fiddle around with small things that are simpler to address, but not worth the candle.

Now I'm not saying that 20mph speed limits in towns are always inappropriate, far from it, rather I'm saying this blanket approach is disproportionate. Norway has the best road safety record in the world - despite relatively high proportions of the kind of roads that present more accidents. What we see there is much more graduated use of lower speed limits in towns. There are 30kmh areas right in the town centre, generally enforced with speed bumps, but more graduated speed limits as you move away from the very centre. Of course it helps to use km so that 40kmh is a commonly used speed limit, presenting graduation between 20mph and 30mph, in this country where you never see 25mph. The Norwegian approach, in a country with a very high quality of life, seems much more proportionate to me.

[lpm]

1. Wales has not gone for a blanket approach. It is doing a temporary blanket approach, followed by revisions to speed limits in coming years. Clearly there are inappropriate limits in some places right now but these will disappear over time.

2. You have failed to include the value of luxury experiences. A house with an attractive view sells for more than an identical house without a view. Likewise a house on a street with quiet sedate traffic gives higher utility to the residents than an identical house on a street with noisy fast traffic.

3. You have failed to include the value of a return to community use of roads, in particular for children. Across decades children were forced indoors by cars and worried parents, and it will probably take decades till we know if perceptions of safer streets will lead to children messing around on bikes etc again. The value of unsupervised unstructured play is hard to determine, but there's a concern that nonstop adult supervision has caused snowflakism.

[El Pollo Diablo]

I thought the non-stop adult supervision was because of teh peedos rather than road safety tbh

[bob sterman]

Now I'm not saying that 20mph speed limits in towns are always inappropriate, far from it, rather I'm saying this blanket approach is disproportionate. Norway has the best road safety record in the world - despite relatively high proportions of the kind of roads that present more accidents. What we see there is much more graduated use of lower speed limits in towns. There are 30kmh areas right in the town centre, generally enforced with speed bumps, but more graduated speed limits as you move away from the very centre. Of course it helps to use km so that 40kmh is a commonly used speed limit, presenting graduation between 20mph and 30mph, in this country where you never see 25mph. The Norwegian approach, in a country with a very high quality of life, seems much more proportionate to me.

A bit like this...

https://www.valeofglamorgan.gov.uk/en/l ... Wales.aspx

Exceptions to the default 20mph limit

On some roads, a 20mph speed limit would not be appropriate or feasible. These roads will be known as exceptions, and the 30mph speed limit can remain. We will follow the statutory Traffic Regulation Orders (TROs) process to make exceptions. The TRO process will also include proposals for 30mph buffer zones. These zones will act as transitions from the higher limit into the new 20mph default limit and additional lengths of roads which considered appropriate to reduce to 20mph by Order for consistency and to simplify signing requirements.

[IvanV]

A bit like this...

https://www.valeofglamorgan.gov.uk/en/l ... Wales.aspx

Exceptions to the default 20mph limit

On some roads, a 20mph speed limit would not be appropriate or feasible. These roads will be known as exceptions, and the 30mph speed limit can remain. We will follow the statutory Traffic Regulation Orders (TROs) process to make exceptions. The TRO process will also include proposals for 30mph buffer zones. These zones will act as transitions from the higher limit into the new 20mph default limit and additional lengths of roads which considered appropriate to reduce to 20mph by Order for consistency and to simplify signing requirements.

No, nothing like that. Wales provides for a small number of exceptions to its near-blanket provision. And if you travel on main suburban arterial routes through Welsh towns and villages, you will discover these 30mph exceptions are uncommon, thinly supplied and for relatively small distances. In Norway, the 20mph section is only a small proportion of the main traffic corridors in towns. Rather you approach the town at 70kmh, here travelling through the scattered outer habitations, and then as the suburbia becomes more continuous, you are reduced in stages through 60kmh, 50kmh, 40kmh as you get into more peopled areas, until you arrive at the 30kmh section with speed bumps in the town centre. When the speed bumps run out, it is 40kmh. It really is completely different.

[dyqik]

I thought the non-stop adult supervision was because of teh peedos rather than road safety tbh

It's due to both, plus others.

Streets are busier than earlier times, even with parked cars and no moving traffic, and there's been a general loss of safe outside space for kids. The number of front gardens paved over for driveways is also part of this loss of space to cars.

[dyqik]

A bit like this...

https://www.valeofglamorgan.gov.uk/en/l ... Wales.aspx

Exceptions to the default 20mph limit

On some roads, a 20mph speed limit would not be appropriate or feasible. These roads will be known as exceptions, and the 30mph speed limit can remain. We will follow the statutory Traffic Regulation Orders (TROs) process to make exceptions. The TRO process will also include proposals for 30mph buffer zones. These zones will act as transitions from the higher limit into the new 20mph default limit and additional lengths of roads which considered appropriate to reduce to 20mph by Order for consistency and to simplify signing requirements.

No, nothing like that. Wales provides for a small number of exceptions to its near-blanket provision. And if you travel on main suburban arterial routes through Welsh towns and villages, you will discover these 30mph exceptions are uncommon, thinly supplied and for relatively small distances. In Norway, the 20mph section is only a small proportion of the main traffic corridors in towns. Rather you approach the town at 70kmh, here travelling through the scattered outer habitations, and then as the suburbia becomes more continuous, you are reduced in stages through 60kmh, 50kmh, 40kmh as you get into more peopled areas, until you arrive at the 30kmh section with speed bumps in the town centre. When the speed bumps run out, it is 40kmh. It really is completely different.

I guess you just ignored LPM's point 1.

[IvanV]

1. Wales has not gone for a blanket approach. It is doing a temporary blanket approach, followed by revisions to speed limits in coming years. Clearly there are inappropriate limits in some places right now but these will disappear over time.

2. You have failed to include the value of luxury experiences. A house with an attractive view sells for more than an identical house without a view. Likewise a house on a street with quiet sedate traffic gives higher utility to the residents than an identical house on a street with noisy fast traffic.

3. You have failed to include the value of a return to community use of roads, in particular for children. Across decades children were forced indoors by cars and worried parents, and it will probably take decades till we know if perceptions of safer streets will lead to children messing around on bikes etc again. The value of unsupervised unstructured play is hard to determine, but there's a concern that nonstop adult supervision has caused snowflakism.

These are all good points.

As for 1, at the moment it is a near-blanket, and I think people don't necessarily believe in its temporality, temporary things have a tendency of becoming permanent. Maybe that was even the idea. My suspicion is that lip service will be paid by a small number of adjustments, which the the politicians can claim gratitude for. At the moment, we don't see a lot of proportionality. The processes make it look like avoiding the 20mph speed limit will be hard. But let us see if what you say really comes to pass. I think it would have been better to invent a more proportionate framework in the first place.

In relation to 2, it is a feature of the British planning system that a planner decides that a cost on one set of people is justified by an advantage to another set of people, but there is never any attempt for the latter set to compensate the former set. Indeed it is illegal to pay someone to remove their objection to a planning application. Some other countries do at least make some attempt to compensate riparian land-owners who suffer large reductions in amenity from development. In France, the approach of build first and pay people off later is seen as presenting much greater ease and speed of development. On the other hand, there is the issue of what do you own. And places where people often claim to own something which the behaviour of others impinges on, seek to restrict the behaviour of others by claiming that ownership, unless paid off, and this can gum up the wheels of the economy and reasonable enjoyment of others. I did some work on railways in Colombia, and the main Colombian freight railway had a huge problem doubling its track, as all sorts of traditional rights appeared that it had to spend ages negotiating compensation for. In Dublin, where the tram ran along the street, a lot of riparian land owners got very nice compensation for having the tram run along the street they fronted, which all took a long time to negotiate and added a lot of cost. That's not considered very usual. So the best way of addressing this is not clear.

But in general, I do think this is a fair point and a bit of attention should be given to it, provided it can be done in a way which doesn't add huge delay and cost to reasonable development, because we have too much of that already.

As for 3, yes, once you get into places that don't go anywhere and are where people live, it is a widespread 30kmh speedlimit in Norway too, and not necessarily completely covered in speed bumps. And that is fair enough. And that is what is happening in London and I think it is proportionate.

At the end of the day, did anyone seek to quantify any of these things, and work out how they balanced out? For a policy that imposes 1.0bn minuites of delay per year on the road-users of Wales, I think they should perhaps have been a bit more quantitative about it.

[lpm]

It's not complicated. Residents in road A campaign for 20 mph in road A, want 30 mph to remain in roads B, C and D. Residents in road B campaign for 20 mph in road B, want 30 mph to remain in roads A, C and D. And so on.

This effect is clearly evidenced in surveys, Facebook groups and pub conversations.

Therefore, the planning system doesn't need to get complicated or gummed up. It's a simple override, listening to residents and ignoring passers through.

[Beaker]

The Welsh analysis suggests an average journey time increase as 50 seconds per journey. I would describe that as trifling.

You could lose a similar amount of time if someone ahead needs to turn right, or the lights change, or it is a bit busy. Adding up everyone’s 50 seconds x a rate per second is useful to compare options, but if we all got there 50 seconds sooner we will not be £110m better off.

My experience of driving through Wales is that in urban environments I am already going slowly, so I’ll happily live with it. Especially as it makes it much safer if I’m on the bike.

ETA link
https://www.gov.wales/evidence-about-20 ... rney-times

[Opti]

Here in Spain the urban restrictions are simply spelt out. 50 km/h if there are 2 lanes in each direction, 30 km/h if 1 lane each way and 20 km/h if a single lane in each direction has pavements next to the road. Pretty rigorously enforced by radar and other means. The vast majority comply and no one really moans.
This may be because the penalty fines can be eye-watering.

[Fishnut]

Road casualties in Wales would have to reduce by about 29% to justify the extra travel time (I used the unrounded figures for that ratio).

Bristol began introducing 20mph speed limits in 2008. Research published in BMJ Injury Prevention in 2020 reported that the reduction was associated with around a 63% reduction in fatalities, though with admittedly very large confidence intervals. They also found a general reduction in injuries since the reduction in speed limit was reduced.

[dyqik]

The Welsh analysis suggests an average journey time increase as 50 seconds per journey. I would describe that as trifling.

You could lose a similar amount of time if someone ahead needs to turn right, or the lights change, or it is a bit busy. Adding up everyone’s 50 seconds x a rate per second is useful to compare options, but if we all got there 50 seconds sooner we will not be £110m better off.

My experience of driving through Wales is that in urban environments I am already going slowly, so I’ll happily live with it. Especially as it makes it much safer if I’m on the bike.

ETA link
https://www.gov.wales/evidence-about-20 ... rney-times

Going slightly further and thinking about the likely distribution of delays.

Time delays under something like about 5 minutes are buried by noise in traffic, phasing of lights, parking, leaving the house, etc.. That's going to cover the vast majority of the power spectrum of journey delays (up to 6 times the mean).

[IvanV]

The Welsh analysis suggests an average journey time increase as 50 seconds per journey. I would describe that as trifling.

The whole point of my post is that we mustn't look at an apparently small number, and call it trifling, without actually thinking what it means. That's why calling it 1.0bn person minutes of extended journey time is a much better way of looking at it, than an average of 50 seconds per journey, which looks like a small number. (And, btw, the Welsh link and the info - 1.0bn person minutes for 1.2bn person journeys - was implicit in my initial post.) The 50 seconds was in the public domain a long time ago. It is only by finding the crucial piece of information, 1.2bn person journeys, that I could quantify what it really meant.

Laughing at a small number - we can call it the fallacy of proof by ridicule - see my recent post here - was the same approach taken by William Happer, Trump's advisor on climate change. He wrote a paper in which he computed exactly the same forcing from CO2 as everyone respectable. And then he went, that's a small number, nothing to see around here. The same logical fallacy I'm trying to explode here. Happer's paper is here, in case you want to check that out.

Of course those 50secs will be distributed, so there will be some who are barely delayed, and some who are delayed much more. Maybe there is a consistent population of urban dwellers and commuters who are being delayed a lot. Couriers who make a lot of short trips in urban areas are seriously pissed off.

And people say, but that 50 seconds disappears into the noise. But if you have a journey of a noisy length of time, and it is now 1 minute longer on average, then you have to set off 1 minute earlier. It doesn't disappear into the noise. And at the end of all that, the Welsh population is spending 1.0bn more person minutes to do what it used to do.

Let's decide whether 1.0bn person minutes really is trifling in comparison to the benefit claimed, rather than just laughing at it, as Happer laughed at the modest amount of climate forcing that the CO2 increase provides.

[nekomatic]

I’m not a transport economist, but if the calculated benefits of a speed limit reduction in urban areas to 20 mph don’t outweigh the calculated costs then we ought to consider whether the benefits of a speed limit increase in urban areas - to 40 mph, say - would outweigh the costs.

If our calculation were to show that the latter was true, and we found that rather unacceptable, then maybe we might conclude that our calculations didn’t actually capture our values very accurately.

[Brightonian]

Very long time ago, somewhere else (not this forum's predecessor), I complained about my train commute being extended some minutes after a convenient exit from a station was closed off, forcing me to go a roundabout route including waiting to cross the road at some lights. Nobody else was interested, I suppose because as a humble pedestrian my time didn't matter.

[IvanV]

I’m not a transport economist, but if the calculated benefits of a speed limit reduction in urban areas to 20 mph don’t outweigh the calculated costs then we ought to consider whether the benefits of a speed limit increase in urban areas - to 40 mph, say - would outweigh the costs.

If our calculation were to show that the latter was true, and we found that rather unacceptable, then maybe we might conclude that our calculations didn’t actually capture our values very accurately.

In a general sense, you are very right, this is the kind of calculation people don't do. Especially they don't calculate if Dr Beeching might be right again today, and whether we should lop off various lighter used bits of our railway system.

It is also the mistake that people make when they say we shouldn't have any new roads anywhere ever, or widenings, with exceptions limited to new housing developments. This is something the Welsh have also said, and I had a rant on the fallacy of it a while ago. But it turns out that they aren't quite as rigid in their application of that policy as their statement of it made it look.

But I haven't set out to suggest that all 20mph speed limits are wrong, rather that not all 20mph speed limits are right. Indeed I didn't even quite say that, I said that they had failed to justify it. 20mph speed limits, I was careful to say, are likely make a lot of sense in the right places. The problem is the near-blanket application of them to nearly all urban roads, and the lack of thought as to whether that is sensible. (Pace lpm's point 1, which I discussed earlier). So I don't think actually the problem you mention really arises as a logical consequence of my argument.

But your point is relevant to what we often see, where a speed limit of 30 mph applies until suddenly it becomes 60mph. This lack of graduation seems unlikely to be optimal to me. In Norway, you see a lot more graduation as towns peter out. Though in Norway, towns frequently peter out in a more gradual way than is common in Britain, where sudden edges of towns are a not uncommon artefact of the planning system.

[Fishnut]

Very long time ago, somewhere else (not this forum's predecessor), I complained about my train commute being extended some minutes after a convenient exit from a station was closed off, forcing me to go a roundabout route including waiting to cross the road at some lights. Nobody else was interested, I suppose because as a humble pedestrian my time didn't matter.

I find that pedestrians are expected to accept inconvenience in a way that road users rarely are. My biggest bugbear is pedestrian crossings where you have to cross to the middle of the road to then wait for lights to change to continue crossing on the other side because we can't possible stop both directions of traffic in one go. The idea that I would want to cross the entire road in one go is apparently unusual and I should be happy to wait 30+ seconds for the lights to change while I wait in the middle of a busy road (and that middle crossing will often require me to S through it because who would want to walk in a straight line?)

Roadworks where they block the pavement and expect you to cross to the other side of the road (assuming there's even pavement on the other side of the road) rather than provide a pedestrian path through also annoy. Even better is when the pavement isn't cut off but they decide that's the best place to stick their 'road work' sign, blocking it anyway.

Once you start actually looking at pedestrian infrastructure you realise how much of it is designed to make you take the long way round. Even things as simple as crossing a side road often means having to go into the side road a few metres then crossing and then walking back up to the main road (if the road isn't fenced at the corner you can take your own line, but if you require the ramps then you are forced to take the longer route).

[IvanV]

Very long time ago, somewhere else (not this forum's predecessor), I complained about my train commute being extended some minutes after a convenient exit from a station was closed off, forcing me to go a roundabout route including waiting to cross the road at some lights. Nobody else was interested, I suppose because as a humble pedestrian my time didn't matter.

Well actually when doing station designs, at least for major stations, railway planners do take very careful account of pedestrian walk routes and walk times, and if they have to make level changes. There are valuations for these timings and inconveniences, which are treated just as anything else. The pedestrian approach to the station is considered important to. There are now grandiose plans around for making improvements to various major stations on such grounds. But of course they depend on lots of money which is short supply.

Though doesn't mean that is what in the end drives the development. Going back to times when money was less of a problem, when Network Rail first suggested doing something like the Shard on top of London Bridge station, based on various station alterations to facilitate the Crossrail expansion, I looked at their pedestrian models. I pointed out that they had sneakily found a way of keeping the longlisted designs with shorter walk times and fewer level changes out of the shortlist, so that they didn't have to do a proper evaluation of them. I suspected some of them might evaluate rather better when you did it carefully. But the "problem" was that the minister was much more likely to be interested in an option with a luvly jubbly large development project, and a till various people could get their fingers in. Anyway, my comments did help get the plan kicked down the road at that point in time. But something similar came back later, got approved, and now we have the Shard and the London Bridge upgrade with its long walk routes.

But in your case, that's the kind of thing they can just do on the fly without a proper evaluation. For operational convenience, they will say. And probably without making any such evaluation. My local station, we did eventually make them see sense and keep entrances both sides open all the time - car park one side, town the other side. The next local station, that I mostly use, has a similar issue, and so entrances both sides are open all the time. But it does make paying to travel between them optional, as my daughter pointed out. But another, larger, local station used to have entrances on both sides of the tracks, and one got permanently closed.

[Gfamily]

The Welsh analysis suggests an average journey time increase as 50 seconds per journey. I would describe that as trifling.

The whole point of my post is that we mustn't look at an apparently small number, and call it trifling, without actually thinking what it means. That's why calling it 1.0bn person minutes of extended journey time is a much better way of looking at it, than an average of 50 seconds per journey, which looks like a small number. (And, btw, the Welsh link and the info - 1.0bn person minutes for 1.2bn person journeys - was implicit in my initial post.) The 50 seconds was in the public domain a long time ago. It is only by finding the crucial piece of information, 1.2bn person journeys, that I could quantify what it really meant.

Tim Harford, in his Radio and Podcast series, often asks "Is that a big number?"
1 billion looks like a big number, but it needs to be scaled against the total number of journey minutes. So, what is the 1 billion against the total number?
Genuine Question.

[dyqik]

The Welsh analysis suggests an average journey time increase as 50 seconds per journey. I would describe that as trifling.

The whole point of my post is that we mustn't look at an apparently small number, and call it trifling, without actually thinking what it means. That's why calling it 1.0bn person minutes of extended journey time is a much better way of looking at it, than an average of 50 seconds per journey, which looks like a small number. (And, btw, the Welsh link and the info - 1.0bn person minutes for 1.2bn person journeys - was implicit in my initial post.) The 50 seconds was in the public domain a long time ago. It is only by finding the crucial piece of information, 1.2bn person journeys, that I could quantify what it really meant.

What we also shouldn't do, though, is to look at a lot of very small numbers, and add them up without thinking about whether noise in the system means that each individual number is meaningless in the vast majority of cases, and whether that noise integrates down when adding the individual small numbers together. In particular, a small change in a number with a high variance may not make any significant difference if each individual doesn't make many trials, and that number is a small fraction of the total experienced by each individual. If the noise doesn't integrate down, then the large sum is equally meaningless. How it integrates down also affects the conclusions, as do the consequence of that number being larger.

A 1 minute delay on a journey that is typically 4-6 minutes (or 40-60 minutes) is not significant to most people who are only making a few journeys a day. They can make that 1 minute up in all sorts of other ways, because people do not optimally use their time, and generally don't want to (optimising has a cost). The people who lose anything by it now being 5-7 minutes are those making a very large number of journeys a day, where the small changes to each journey do integrate (this is essentially the same argument as the threshold effect in e.g. radiation exposure, or infectious disease exposure).

CO2 emissions integrate, by the way, which is why this argument doesn't work there. An extra 10% emissions from you is additive with an extra 10% emissions from me, in a way that an extra 10% time spent driving to the shops for you does not add to an extra 10% of time spent driving to shops for me (both do add to the total number of cars on the road though. The effect on emissions is more complex, as driving faster and accelerating harder produces more emissions than extra time spent driving at a slower speed).

[dyqik]

You should probably also consider the benefits to drivers of driving slower. Firstly, generally it costs less per mile - in fuel, maintenance, and in insurance rates (to the extent that everyone is driving more slowly, and thus safely, or with lower repair costs in an accident). Secondly, it's usually less stressful. Thirdly, it gives more mental space for non-driving thoughts - I may well have a very valuable idea in the extra couple of minutes on my journey that's undertaken at lower mental load.

[Fishnut]

You should probably also consider the benefits to drivers of driving slower. Firstly, generally it costs less per mile - in fuel, maintenance, and in insurance rates (to the extent that everyone is driving more slowly, and thus safely, or with lower repair costs in an accident). Secondly, it's usually less stressful. Thirdly, it gives more mental space for non-driving thoughts - I may well have a very valuable idea in the extra couple of minutes on my journey that's undertaken at lower mental load.

Also means they have a greater ability to check their surroundings and stop in time when the unexpected happens. A near miss with a kid is far less stressful than a hit.

[dyqik]

You should probably also consider the benefits to drivers of driving slower. Firstly, generally it costs less per mile - in fuel, maintenance, and in insurance rates (to the extent that everyone is driving more slowly, and thus safely, or with lower repair costs in an accident). Secondly, it's usually less stressful. Thirdly, it gives more mental space for non-driving thoughts - I may well have a very valuable idea in the extra couple of minutes on my journey that's undertaken at lower mental load.

Also means they have a greater ability to check their surroundings and stop in time when the unexpected happens. A near miss with a kid is far less stressful than a hit.

Another benefit to drivers is that slower speed limits increase road capacity, and thus reduce variability in journey times for roads that are near capacity. This may well result in drivers being able to leave later to arrive on time for appointments, work etc., even if the average journey time has increased.

[nekomatic]

In a general sense, you are very right, this is the kind of calculation people don't do. Especially they don't calculate if Dr Beeching might be right again today, and whether we should lop off various lighter used bits of our railway system.

I applied the same logic when I heard people arguing that HS2 was bad for Manchester, because a better connection to London would take money and jobs away from the former in favour of the latter. Little did I imagine that Avanti would try the experiment.

It’s obviously hard to quantify how people really feel about something - especially when they probably don’t know themselves - but I think it’s worth attempting.