How to make safety decisions - Wales urban 20mph

[Sciolus]

I don't find driving at 20mph low stress, I find it high stress. I find it hard to manage the car to maintain at that speed, the accelerator isn't very forgiving around there so my speed is never right. And I'm taking my eyes off the road to look at the speed all the time. The speedo isn't very accurate down there either. For that reason, there is actually more mental load.

Since Neko mentions it but doesn't actually respond: Any competent British driver can drive at a steady 30 mph without looking at the speedo more than occasionally. This is from frequent practice. I'm sure if Ivan spent a bit more time driving at 20 he soon would find himself not relying on the speedo and it will become second nature.

[TopBadger]

And you can always drive in a higher ratio gear to give more throttle control... indeed speed awareness courses advise driving in 3rd gear when doing 30mph.

[IvanV]

I'm rather puzzled that Ivan started this thread without thinking to type "Wales 20 mph regulatory impact assessment" into his favourite search engine. He could have saved himself an envelope.

The RIA makes clear that, although there are large uncertainties because nothing like this has been done in the UK before, the monetised cost of additional journey times (central £6.4b over 30 years) far exceeds the monetised benefits. But the measure supports a wide range of policies (Section 2.2), with the particular aims to:

2.2.4.1 Reduce injuries on the road network.
2.2.4.2 Encourage a change in travel behaviour, with people feeling confident, safe, and secure enough to increase their use of active travel modes
2.2.4.3 Improve the environment and economy of local communities by reducing the negative externalities associated with vehicle use.

Regarding the matter of whether lots of individual minutes can be added up linearly to get a sensible large number, it says:

It should also be noted that this dis-benefit accrues from the application of standard
constant values of time over which there is active professional debate 80 . There has
also long been academic debate around the application to small time savings 81
where the benefit of small time savings may be perceived to be small or negligible
relative to larger time savings. This being said current TAG guidance stipulates the
use of a constant value of time regardless of the sign or size of the time saving (see
further discussion in section 6.5).

I am grateful to you for correcting this unusual piece of incompetence relative to my usual dogged persistence to locate information if it can be found. You are right, I didn't think to google impact assessment. Often impact assessments are full of unquantified gobbledegook, and the reports with the quantifications - if it has been done - have other names. But in this case I would have hit the jackpot with the first hit if I had googled just the word assessment alongside Wales 20mph. I was trying to google for the detailed data that would underlie that quantification, and that didn't work. At least not until the recent hoohaa where they were made to publish more detail of the data underlying that quantification.

So the safety engineer called this "trifling" even though he, too, should have been aware of just what I discovered with my back of the envelope, that by the standard TAG assessment this policy quantified with a large disbenefit. By these methods, the time lost was not "trifling" at all.

I think it is good nevertheless that I have provoked a lively debate around these complex issues. Many very good points have been made, and it is always good to learn what objections people make, and consider what validity there is in these objections, for there usually is some.

I did come across a case earlier in my career where it was evident that time had quite different unit values at different quantities. This was the case of how long it took rail freight to arrive. With one important exception, most freight movers expected their freight to arrive at a rather random time, and had devised their arrangements to receive it so that it didn't really matter when it arrived within a rather broad window. But once it was outside that broad window - the depot had closed for the day - the cost of the train turning up outside that window was large indeed. And sometimes they had to put on trucks at short notice to bring in supplies which hadn't arrived by rail. The exception to this was power station coal, where if the coal arrived in a narrow window, it could be dropped straight into the furnace, whereas if it arrived outside that, it would have to be dropped onto a coal heap and moved later.

Generally speaking, with travel time savings for people, these are usually thought more robust for smaller rather than larger quantities of time. Whether a journey takes 10 rather than 8 hours is very different proposition per minute from whether a journey takes 40 rather than 45 minutes. The quantifications we see in TAG are quantified on cases related to the latter. There is some indication that journey time extensions - of relatively short journeys - in the range of 20-40 minutes are particularly valuable, which is why the HS1 domestic services to Kent, and Heathrow Express, can charge such a nice premium for their services.

If there are material numbers of people who are losing 5 or 10 minutes a day from this, I would have no hesitation in applying the usual value of time to them. This is the kind of journey time change that we tend to think these values are good for. And indeed for somewhat larger amounts too. If we did have a schedule of values for different amounts of time, we might find that we are undervaluing the 10 minute case and overvaluing the 30 seconds cases. Maybe applying the average value to the total amount is fine. Maybe the shape of the distribution and schedule of values is such that it isn't. It doesn't seem to be easy to know.

[El Pollo Diablo]

I'm rather puzzled that Ivan started this thread without thinking to type "Wales 20 mph regulatory impact assessment" into his favourite search engine. He could have saved himself an envelope.

The RIA makes clear that, although there are large uncertainties because nothing like this has been done in the UK before, the monetised cost of additional journey times (central £6.4b over 30 years) far exceeds the monetised benefits. But the measure supports a wide range of policies (Section 2.2), with the particular aims to:

2.2.4.1 Reduce injuries on the road network.
2.2.4.2 Encourage a change in travel behaviour, with people feeling confident, safe, and secure enough to increase their use of active travel modes
2.2.4.3 Improve the environment and economy of local communities by reducing the negative externalities associated with vehicle use.

Regarding the matter of whether lots of individual minutes can be added up linearly to get a sensible large number, it says:

It should also be noted that this dis-benefit accrues from the application of standard constant values of time over which there is active professional debate 80 . There has also long been academic debate around the application to small time savings 81 where the benefit of small time savings may be perceived to be small or negligible relative to larger time savings. This being said current TAG guidance stipulates the use of a constant value of time regardless of the sign or size of the time saving (see further discussion in section 6.5).

I was going to enquire along these lines, without wishing to contribute to a pile-on. Ivan, was your analysis for just one year or a longer period? Because discounting would normally apply to these scenarios, and as TAG & the HMT Green Book make clear, the discount rate for all non-safety costs & benefits is 3.5% pa for the first 30 years whereas for safety-related costs & benefits, it's 1.5% pa (up to year 30). Sciolus' link above (p18) shows that the analysis follows this. Over time, this (deliberate) bias would be expected to see the safety benefits maintain their value better than any costs, tipping the scales in favour of the benefits.

That said, even with that bias, the report isn't exactly glowing about the prospects of the policy, but it also hasn't had the traffic impacts worked out in detail.

ETA: There's debate, probably outside HMT, about whether the discount rate should be 3.5% given that around 60% of that is due to estimates of historic wealth growth in the post-WW2 period, something which has singularly failed to replicate itself in the last twenty years.

[dyqik]

The overall point is that saving a couple of minutes a day in journeys doesn't have any significant value to most people, and you can't add up insignificant value to lots of people to get significant value - because it's always distributed.

If a couple of minutes a day don't have any significant value to you - would you mind doing a couple of minutes a day of work for me? Unpaid of course. Yes you'll be doing the equivalent of approximately 2 days work for me over a year (about 12 hours). But that doesn't have any significant value right? I could make a bundle if I could get lots of people to do the same!

Or perhaps, if you have a car, you wouldn't mind if I switched it for one with a dodgy engine that takes a couple of minutes to start each morning? Again - 12 hours a year you'd be sitting their groaning because it wont start - but that time has no value right?

You've missed the point entirely here.

Would you pay me to work for you for 2 minutes a day? Would that be worth your time?

How about 20 minutes every two weeks?

Or 1 day a year?

Similarly, does delaying 1.2 billion people by 50 seconds have the same impact as delaying 12 million people by an hour and a twenty minutes?

You can't just add up tiny increments of time to a big number that you assign fixed value to, if those tiny increments are not adjacent. Small delays to many people do not integrate.

By the way, I used to drive an MGB as my only car, in the 2010s. So yes, I will happily put up with a car that takes a minute or two more to start per day. What I won't put up with is one that I lose access to for a week every couple of years because I can't fix it.

[dyqik]

Similarly, every day I wait at the station for roughly between five to ten minutes for my train.

Should I leave the house two minutes later because that time is so valuable?

[dyqik]

The overall point is that saving a couple of minutes a day in journeys doesn't have any significant value to most people, and you can't add up insignificant value to lots of people to get significant value - because it's always distributed.

If a couple of minutes a day don't have any significant value to you - would you mind doing a couple of minutes a day of work for me? Unpaid of course. Yes you'll be doing the equivalent of approximately 2 days work for me over a year (about 12 hours). But that doesn't have any significant value right? I could make a bundle if I could get lots of people to do the same!

Or perhaps, if you have a car, you wouldn't mind if I switched it for one with a dodgy engine that takes a couple of minutes to start each morning? Again - 12 hours a year you'd be sitting their groaning because it wont start - but that time has no value right?

You've missed the point entirely here.

Would you pay me to work for you for 2 minutes a day? Would that be worth your time?

How about 20 minutes every two weeks?

Or 1 day a year?

Similarly, does delaying 1.2 billion people by 50 seconds have the same impact as delaying 12 million people by an hour and a twenty minutes?

You can't just add up tiny increments of time to a big number that you assign fixed value to, if those tiny increments are not adjacent. Small delays to many people do not integrate.

That is, no, I wouldn't work for you for free for two minutes a day, because it would not produce any value. I would volunteer for an entire day once a year though.

[dyqik]

I'm just amazed at all these people who are able to travel to such tight schedules that even a minute's difference will be an inconvenience. Traffic lights are always on their side, a parking space is always available right where they want it, nothing unexpected happens that means their journey ever fluctuates and makes having a buffer of a couple of minutes necessary.

There are no buses, cyclists, pedestrians using crossings, etc. as they drive their spherical cars along the frictionless road in a vacuum tube.

[dyqik]

The name for the fallacy above when applied to project management, by the way, is "The Mythical Man Month".

[dyqik]

I'm just amazed at all these people who are able to travel to such tight schedules that even a minute's difference will be an inconvenience. Traffic lights are always on their side, a parking space is always available right where they want it, nothing unexpected happens that means their journey ever fluctuates and makes having a buffer of a couple of minutes necessary.

There are no buses, cyclists, pedestrians using crossings, etc. as they drive their spherical cars along the frictionless road in a vacuum tube.

Also, they never let anyone out of a turning ahead of them, never give way, never hold a door for someone, never stop to chat, make tea with water that hasn't quite boiled yet, never post on pointless Internet forums.

[Sciolus]

I am grateful to you for correcting this unusual piece of incompetence relative to my usual dogged persistence to locate information if it can be found. You are right, I didn't think to google impact assessment. Often impact assessments are full of unquantified gobbledegook, and the reports with the quantifications - if it has been done - have other names. But in this case I would have hit the jackpot with the first hit if I had googled just the word assessment alongside Wales 20mph. I was trying to google for the detailed data that would underlie that quantification, and that didn't work. At least not until the recent hoohaa where they were made to publish more detail of the data underlying that quantification.

I hope I didn't come across as snarky. I was genuinely a bit surprised, because you know far more about how this works than most people hereabouts (certainly including me). I generally agree with the thrust of your OP.

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.

An important question. The way I deal with big national-scale numbers is simply to work on a per-capita basis. The population of Wales is 3.1m, so we're talking a cost somewhere in the region of £70 per person per year (based on the RIA figure and not bothering to correct for discounting, it's all massively uncertain, and if someone want to do the calculation properly please do). That seems like a number that not many people would want to pay for the perceived benefits.

BUT.

This whole discussion is framed arse-about-face.

People are talking about imposing costs on motorists in order to provide benefits to third parties. In fact, we are redistributing a fraction of the costs that motorists impose on third parties back onto the motorists. That seems rather harder to argue against (without going "waaaah, road tax!!11!").

[dyqik]

£70 a year is about what people waste on excess housing in order to store a couple of boxes of old cables, some clothes that don't fit them any more, and to have an extra layer of paint on the walls.

It's way way less than what they waste on owning a newer car that only gets used an hour a day.

[dyqik]

So, how much should we charge drivers in single occupancy vehicles for causing excess traffic and delaying buses and people car pooling?

[El Pollo Diablo]

we're talking a cost somewhere in the region of £70 per person per year (based on the RIA figure and not bothering to correct for discounting, it's all massively uncertain, and if someone want to do the calculation properly please do).

Costs discount at 3.5% per year for the first 30 years, so for a constant annual cost over 30 years the NPV is 19.4 times the cost (so divide by 19.4). Assuming you originally divided by 30, that would be around £108pp pa.

Safety costs discount at 1.5% per year for the first 30 years, so for that the NPV is 25 times the cost.

[IvanV]

This whole discussion is framed arse-about-face.

People are talking about imposing costs on motorists in order to provide benefits to third parties. In fact, we are redistributing a fraction of the costs that motorists impose on third parties back onto the motorists. That seems rather harder to argue against (without going "waaaah, road tax!!11!").

This is a lovely point. Thank you. I'll spend some time digesting it.

I am grateful to you for correcting this unusual piece of incompetence ...

I hope I didn't come across as snarky. I was genuinely a bit surprised, because you know far more about how this works than most people hereabouts (certainly including me). I generally agree with the thrust of your OP.

No, and I was genuinely mortified at my stupidity, especially when "impact assessment" had gone across my mind at one point. I was probably feeling smug from having googled an obvious thing related to something my work colleagues had been writing about last week, only to discover some information that was fatally inconvenient for what they had been writing.

[lpm]

People are talking about imposing costs on motorists in order to provide benefits to third parties. In fact, we are redistributing a fraction of the costs that motorists impose on third parties back onto the motorists. That seems rather harder to argue against (without going "waaaah, road tax!!11!").

But the motorists and the third parties are almost the same people.

I don't know how we'd get the % people who live on a street that is now 20 mph. But as 20 mph covers cities and rural villages it is going to be very high.

The actual question is would you pay 1 minute a day in return for quieter and less threatening traffic outside your home?

[IvanV]

But the motorists and the third parties are almost the same people.

About 10% of traffic is work-related - freight, tradesmen, etc. And in the value assessment, work-related traffic has a much higher value per minute. And the value in this work-related traffic comes from the business, not the driver - though some may be self-employed. So maybe about 40-50% of the value is not in the people driving/travelling, but in their business.

And then in Wales 26% of households have no access to a car.

And there will be a proportion of non-Welsh residents driving through.

So there's a large overlap, but different enough it isn't "almost the same".

[Grumble]

But the motorists and the third parties are almost the same people.

About 10% of traffic is work-related - freight, tradesmen, etc. And in the value assessment, work-related traffic has a much higher value per minute. And the value in this work-related traffic comes from the business, not the driver - though some may be self-employed. So maybe about 40-50% of the value is not in the people driving/travelling, but in their business.

And then in Wales 26% of households have no access to a car.

And there will be a proportion of non-Welsh residents driving through.

So there's a large overlap, but different enough it isn't "almost the same".

In more rural parts of Wales the proportion of non-resident drivers will be quite high during the summer months.

[IvanV]

The actual question is would you pay 1 minute a day in return for quieter and less threatening traffic outside your home?

Of course the cost in terms of minutes day varies. As does the benefit as some people live in quiet cul-de-sacs, etc. But it's similar to the nice comment on "I'd like you to drive slowly past my house while I drive fast past yours."

It is an interesting question whether we would each agree to drive slowly, or faster. I really have no idea what people in general would think, I'm far from an average specimen.

[dyqik]

The actual question is would you pay 1 minute a day in return for quieter and less threatening traffic outside your home?

Of course the cost in terms of minutes day varies. As does the benefit as some people live in quiet cul-de-sacs, etc. But it's similar to the nice comment on "I'd like you to drive slowly past my house while I drive fast past yours."

As I pointed out upthread, it doesn't vary that much - the delays are not proportional to the length of the journey once you get over some distance, and so the maximum cost is pretty tightly constrained.

[IvanV]

As I pointed out upthread, it doesn't vary that much - the delays are not proportional to the length of the journey once you get over some distance, and so the maximum cost is pretty tightly constrained.

Wales isn't actually very well supplied with the kind of segregated major roads that by-pass settlements. There's a lot of A roads, and certainly most of the secondary roads, that are through-every-village-and-town. My general experience of travelling in Wales is a lot of time on such roads. In my last trip in Wales, I did exit on the M4, but the 30miles to get there along an A-road that was through-every-town-and-village. And even on some by-pass-like edge-of-town bits of roads as I approached the motorway, even those were 20mph. But between the A55 racetrack along the north coast, and the segregated highways in the conurbations of the south, there's little like a segregated highway. The Welsh sections of the A5, the A44, the A40 (east of where it becomes the M4-in-all-but-name), the A483, notable main roads between north coast and the southern conurbations, these are through-every-village non-segregated roads for much of their Welsh length.

[lpm]

The actual question is would you pay 1 minute a day in return for quieter and less threatening traffic outside your home?

Of course the cost in terms of minutes day varies. As does the benefit as some people live in quiet cul-de-sacs, etc. But it's similar to the nice comment on "I'd like you to drive slowly past my house while I drive fast past yours."

It is an interesting question whether we would each agree to drive slowly, or faster. I really have no idea what people in general would think, I'm far from an average specimen.

We don't need to ask people, because we already vote with our money. Collectively we pay a premium to live in a quiet cul-de-sac. And pay a discount to live on a busy road.

Overnight hundreds of thousands of homes were promoted to sedate residential neighbourhood.

[dyqik]

As I pointed out upthread, it doesn't vary that much - the delays are not proportional to the length of the journey once you get over some distance, and so the maximum cost is pretty tightly constrained.

Wales isn't actually very well supplied with the kind of segregated major roads that by-pass settlements. There's a lot of A roads, and certainly most of the secondary roads, that are through-every-village-and-town. My general experience of travelling in Wales is a lot of time on such roads. In my last trip in Wales, I did exit on the M4, but the 30miles to get there along an A-road that was through-every-town-and-village. And even on some by-pass-like edge-of-town bits of roads as I approached the motorway, even those were 20mph. But between the A55 racetrack along the north coast, and the segregated highways in the conurbations of the south, there's little like a segregated highway. The Welsh sections of the A5, the A44, the A40 (east of where it becomes the M4-in-all-but-name), the A483, notable main roads between north coast and the southern conurbations, these are through-every-village non-segregated roads for much of their Welsh length.

Arterial roads like those are exactly the kind of roads that are intended to be, and will be, excepted from this.

[IvanV]

Arterial roads like those are exactly the kind of roads that are intended to be, and will be, excepted from this.

So I'm told. I go to Wales a lot and shall see how much "will be" actually happens.

[IvanV]

The actual question is would you pay 1 minute a day in return for quieter and less threatening traffic outside your home?

It is an interesting question whether we would each agree to drive slowly, or faster. I really have no idea what people in general would think, I'm far from an average specimen.

We don't need to ask people, because we already vote with our money. Collectively we pay a premium to live in a quiet cul-de-sac. And pay a discount to live on a busy road.

We know they also put a value on being able to drive past fast past your house.

We can do this rationally. That £4.6bn cost averages something like £3000 per household in Wales, where the average house (outside the social rented sector) is worth £213,000. What is the increase in value from this reduction in speeding past houses, remembering only a selection will experience a material effect. Unfortunately this kind of question tends to be difficult to answer, but we can at least pose it. It isn't obvious to me which way this will lean.

But it sounds like the kind of question that, individually, people won't address in that very rational way. So if you ask someone, who does experience that genuine potential dichotomy, a car owner whose house is located where the speed limit is now reduced to 20mph, what would you rather have, the ability to drive faster past other people's houses, or people driving slower past yours, I wonder what they would say. It might be interesting.