Claim: Denser development increases congestion

TheNZThrower

Active Member
According to a Cato Institute article by a Randal O'Toole, building denser increases congestion and thus CO2 Emissions:
A paper by transportation engineer Kara Kockelman (who was also on the TRB committee) and colleagues at the University of Texas reviews alternative means of reducing greenhouse gas emissions... the paper says, compact development and transit improvements could both substantially increase emissions rather than reduce them—the first by increasing congestion (which leads cars to emit more pollution) and the second because transit construction and operations both emit substantial amounts of greenhouse gases.
This here is the paper by Kockelman. I've noticed a little disparity:
Over the longer term, requirements for better building design, particularly high R-value insulation, a shift toward multi-family structures and smaller dwelling units, use and re-use of lower-energy building materials, and more compact urban arrangements will bear more significant and enduring savings.
So Kockelman appear to say that multi-family housing does bear significant energy savings over the long term.

I'm currently tired, and the paper is long. So if you've got more leads, please reply.
 
You really need to break this down per person to have some comparison.

Article:
One of the biggest differences in carbon footprint is between rural and urban dwellers. Travel and size of homes are the main reasons why rural populations can emit twice as much carbon as their urban counterparts

New York City has an average carbon footprint of 32.6 metric carbon equivalent tons (MCET). People in New York produce only 5 and 7 MCET through travel and housing, since they use public transport or travel short distances. Houses are typical smaller and need less energy to warm.

On the other hand, people in Giltner in Hamilton County in Nebraska produce an average of 65.3 MCET, with the maximum source being transport which is responsible for 23 MCET. Housing is also 22 MCET. Transport is four times and housing is three times more polluting in Giltner than in New York.

Contributions from food, goods, and services are nearly the same in both rural and urban areas.

Cities have a lower carbon footprint only where the people live in smaller houses in crowded cities, like New York City, and commute shorter distances. So New York City has a lower carbon footprint of 32.6 MCET than Denver, CO, where people have an average carbon footprint of 62.1 MCET. This is because in Denver, average houses are bigger, contributing 18 MCET (2.5 times more than New York). The city of Denver is widespread so their transport carbon footprint at 23 MCET is four times more than New York notes a Live Science report.

New York's suburbs are three times as polluting as the national average and more than even some rural areas like Giltner. Not surprisingly, a scientific study reported in Environmental Science and Technology in 2014 showed that suburbs produce approximately 50% of the total U.S. carbon emissions.

Sprawl is bad for the environment. Big houses (which the Cato institute cites as an advantage of low-density building) are bad for the environment.
 
The drawback of a more crowded community can be related to the percent of auto-time that is spent waiting in congestion or at a red light, that is, the time that the engine is running but there is no significant forward motion. In more widely spaced places there is generally less pollution per mile, but more miles. It's a trade-off that doesn't lend itself too well to generalities.
 
The drawback of a more crowded community can be related to the percent of auto-time that is spent waiting in congestion or at a red light, that is, the time that the engine is running but there is no significant forward motion. In more widely spaced places there is generally less pollution per mile, but more miles. It's a trade-off that doesn't lend itself too well to generalities.
this'll go away completely with the transition to electric vehicles that don't need to idle.
 
Going back to the article, Randal claims that a report titled 'Moving Cooler' by the Urban Land Institute backs up his claim that increased density = increased congestion:
Using census data, Moving Cooler estimated that increasing densities from an average of 3,000 people per square mile by an additional 133 percent to an average of 7,000 people would reduce per capita driving by less than 15 percent. That many more people driving 15-percent less each still means a 100-percent increase in total vehicle miles of travel. Since compact-city planners would oppose any new highways to accommodate that travel, there would obviously be a huge increase in congestion.
Looking at Moving Cooler, I ctrl+F'ed for 7,000, per square mile, 15 percent and other terms Randal uses. I found nothing resembling his estimates present. I was only able to find that the figures pertaining to 15% relate to GHG reductions and gallons of fuel saved. As it is a long report, I may have missed out on some things. So if you found out that the report did say what Randal said, prove me wrong.
 
That many more people driving 15-percent less each still means a 100-percent increase in total vehicle miles of travel.
Content from External Source
That's a fallacy, though. These people who moved to that city now don't drive someplace else, resulting in a greater reduction there.

Which is why, if you consider the effects of congestion, the per-capita comparison is important.

Referencing my previous post, someone who moves from Denver to NYC reduces their carbon footprint (on average), even though they're now contributing to congestion in New York.

"More people=more emissions" is kindergarten logic if you don't consider where these people would otherwise be.
 
That many more people driving 15-percent less each still means a 100-percent increase in total vehicle miles of travel.
Content from External Source
That's a fallacy, though. These people who moved to that city now don't drive someplace else, resulting in a greater reduction there.

Which is why, if you consider the effects of congestion, the per-capita comparison is important.

Referencing my previous post, someone who moves from Denver to NYC reduces their carbon footprint (on average), even though they're now contributing to congestion in New York.

"More people=more emissions" is kindergarten logic if you don't consider where these people would otherwise be.
So are you saying that the extra 4,000 people who move there will no longer drive or drive less?

And are you claiming that moving from Denver to NYC tends to reduce a person's carbon footprint via driving?
 
So are you saying that the extra 4,000 people who move there will no longer drive or drive less?
On average, yes.
And are you claiming that moving from Denver to NYC tends to reduce a person's carbon footprint via driving?
And via heating, on average, yes.

You can always pick individuals for which this is not true, e.g. someone who lived in mom's basement in Denver and took a job in NY as a delivery driver will increase carbon emissions, but the carbon footprint of the 4000 people together should drop significantly.
 
On average, yes.

And via heating, on average, yes.

You can always pick individuals for which this is not true, e.g. someone who lived in mom's basement in Denver and took a job in NY as a delivery driver will increase carbon emissions, but the carbon footprint of the 4000 people together should drop significantly.
So Randal claims that:
  1. Increasing density from 3000 to 7000 people per sq. mi decreases driving by 15% per person
  2. If the prior 3000 people drove (let's say) 50 miles per day, then the current miles travelled per capita is 0.85*50
  3. The current miles travelled per capita is 42.5
  4. 3000*50 = 150000 miles
  5. 7000*42.5 = 297500 miles
  6. 297500>150000
  7. Ergo the number of miles driven in total is still greater
  8. This therefore leads to greater congestion in the absence of wider roads
Your counterargument is that the extra 4000 people moving to make a given area denser likely won't drive or don't drive as much as the 3,000 people who lived there prior, ergo the increase in the amount of miles driven is lower than what Randal suggests? Right?

I'm just confused as Randal's argument about increasing the total miles driven leading to more congestion is intuitive to me, and your response seems so out of the blue. Can you explain specifically where Randal get's it wrong in the syllogism? Sorry if I sound like an idiot.

P.S. Did you find anything from Kockelman's paper that resembled what Randal claimed she said?
 
I'm just confused as Randal's argument about increasing the total miles driven leading to more congestion is intuitive to me, and your response seems so out of the blue.
Your title focuses on congestion, but your OP relates this to "the paper says, compact development and transit improvements could both substantially increase emissions" and "significant energy savings". These are not really much related to congestion. The implied argument, 'dense environment have more congestion, therefore they waste energy and kill the planet', is a fallacy, as the MCET statistics I cited show.

Congestion is a function of traffic infrastructure and traffic planning. Once rush-hour traffic has slowed to a crawl, it can't become more congested, no matter how many more people you add. Travelers will switch their habits and walk, bike, or use public transport; these switchovers can be encouraged by urban and traffic planning, i.e. rail systems, bus lanes, bicycle infrastructure etc. European cities (and NYC) do this as a matter of course. Note that these alternate means of transportation are all more energy-efficient than cars.

Cities that do this can grow and not become more congested.

So Randal claims that:
  1. Increasing density from 3000 to 7000 people per sq. mi decreases driving by 15% per person
  2. If the prior 3000 people drove (let's say) 50 miles per day, then the current miles travelled per capita is 0.85*50
  3. The current miles travelled per capita is 42.5
  4. 3000*50 = 150000 miles
  5. 7000*42.5 = 297500 miles
  6. 297500>150000
  7. Ergo the number of miles driven in total is still greater
  8. This therefore leads to greater congestion in the absence of wider roads
Your counterargument is that the extra 4000 people moving to make a given area denser likely won't drive or don't drive as much as the 3,000 people who lived there prior, ergo the increase in the amount of miles driven is lower than what Randal suggests? Right?
My counterargument is that the 4000 would also drive if they lived elsewhere.

3000+4000 *50 = 150000 350 000 miles
7000*42.5 = 7000*42.5 = 297500 miles miles

297 500 < 350 000
ergo, less energy used for traffic

Also, a reduction of only 15% seems low, given the numbers I cited.
 
I've encountered an article from the Victoria Transport Policy Institute by Todd Litman who argues against density=congestion as follows:
How congestion is measured has a major effect on how land use is considered to affect congestion. For example, compact, multi-modal cities such as New York, Boston and Philadelphia tend to have more intense congestion (greater peak-period speed reductions), but lower congestion costs (fewer annual hours of delay per capita) due to lower auto mode shares and shorter trip lengths, which reduces congestion exposure (the amount residents must drive during peak periods). More dispersed, automobile oriented cities such as Houston, Atlanta and Detroit tend to have less intense congestion but greater congestion costs. Compact cities rank worse if evaluated by congestion intensity indicators such as the Travel Time Index (TTI) but better if evaluated by congestion costs, as shown in Table 11.
Here is his table:
Screen Shot 2022-09-26 at 1.58.31 pm.png
As you can tell in the heading, the source for the data is apparently from the 2013 Urban Mobility Report by the Texas Transportation Institute. I can't find the 2013 edition so I will use 2020 data instead from their 2021 report:
Screen Shot 2022-09-26 at 2.07.32 pm.png
I don't know what it was like in 2013, but it appears that the multimodal cities cited by Litman actually rank higher in both travel time index and hours of delay per commuter. Overall, there doesn't seem to be a significant correlation between density and congestion (as determined by TTI & delay per hour per commuter) as many of the most congested cities are also car dependent as well. However, there does appear to be a correlation if we assess the city ranking in 2019. In regards to TTI in both 2019 & 2020, there appears to be no correlation between density and congestion.

So I don't know about you, but it does seem that Litman's statistics are iffy.
 
you forgot what lockdowns and work-from-home did to traffic?

it's not Litman's data that is "iffy"
I did include 2019 data in the table too.

But I also found that the 2013 data available from the TTI doesn't quite match either. According to the Bureau of Transportation Statistics that compiled data from the latest 2021 TTI report, the annual hours of delays per commuter of Boston, New York and Philadelphia are 70, 84 and 54 respectively. Compare this to Litman's data again:
Screen Shot 2022-09-26 at 1.58.31 pm.png
Now I don't know where Litman got his data, but it seems quite off from the actual figures given.

Though the BTS does give this caveat with the data compiled:
Methodology and data sources have been changed in 2019 and were applied retroactively to past years; these figures are not comparable to those in past editions of NTS [national transportation statistics].
 
Wendell Cox, another suburbanist in the vein of O'Toole, claims that higher density development are correlated with more pollution per vehicle distance travelled (VMT/VKT), as well as more congestion per VKT and more VKT per sq. km (or VMT per sq. mi):
1666287834565.png
1666288142906.png

Screen Shot 2022-10-21 at 1.49.36 am.png
The international data reveals virtually the same dynamic. The higher densities of urban areas outside the United States are associated with much higher traffic volumes per square mile (despite their usually far superior transit systems). The highest density urban areas have traffic volume levels 2.5 times that of the lowest (Table). But there is more. Because of the slower speeds associated with higher traffic volumes, the time spent in traffic is even greater. Vehicle hours per square mile are five times as high among the highest density urban areas compared to the lowest.
From this he extrapolates that denser cities have worse congestion and air quality overall:
The fundamental problem is that as population densities rise, vehicle use also rises. Perhaps the most significant research was performed for the Federal Highway Administration, which found, generally, that... if an area experiences an increase of 100 percent in population per square mile, vehicle miles per square mile can be expected to increase 80 percent...

This more intense exposure to air pollution can have negative health impacts (even as overall air pollution levels are falling).
My best counterargument: Let's assume that denser population centres have higher emissions and congestion per VKT and higher VKT per area. This doesn't mean you can therefore extrapolate that denser areas have more congestion or worse air quality overall, especially since any car trips that are made there don't tend to be as long as you don't need to drive as far. What matters is not pollutants per VKT, but pollutants per capita and the overall pollutant level of a city. Likewise, what matters is congestion per capita or per driver than congestion per VKT or VKT per sq. km.

To quote the abstract of a paper by economists Glaeser and Kahn:
Cities generally have significantly lower emissions than suburban areas, and the city-suburb gap is particularly large in older areas, like New York.
 
On top of what Cox has said, I would like to state that from his data, I could extrapolate that while it is reasonable to conclude that the per capita pollutant emissions in a denser area may be lower, the increased VKT per km2. compared to a comparably populated suburban area could mean that the pollutants per area are greater and more concentrated, which can mean worse air quality per area.
 
all of these feel like super obvious claims to me, the question is, why pick/single out these metrics specifically?

Inner city air quality will improve in the coming decade(s) as more vehicles are going to be electric. The real problem is energy expenditure per capita.
 
all of these feel like super obvious claims to me, the question is, why pick/single out these metrics specifically?

Inner city air quality will improve in the coming decade(s) as more vehicles are going to be electric. The real problem is energy expenditure per capita.
Just wanted to summarise Cox's basic claims and then comment on whether his extrapolations are valid.
 
My counterargument is that the 4000 would also drive if they lived elsewhere.

3000+4000 *50 = 150000 350 000 miles
7000*42.5 = 7000*42.5 = 297500 miles miles

297 500 < 350 000
ergo, less energy used for traffic

Also, a reduction of only 15% seems low, given the numbers I cited.
I think what Randal's point is is that even if less people drive in total and even if there are less VKT in total, the increased population density in a city or neighbourhood can cause congestion to worsen in that specific neighbourhood or city as it leads to an increase in VKT/VMT in the area.
 
Just wanted to summarise Cox's basic claims and then comment on whether his extrapolations are valid.
Yes, I understand that.

I'm trying to point out that people can use valid data to distract from the issues that actually matter; it's a way to make a bunk argument that can't be debunked by looking at the details (though it's obviously a good idea to do that anyway), but rather by pointing out that it's of the "rearranging deck chairs on the Titanic" variety.

(If I can walk to the nearest grocery store and take the subway to work, I don't care so much that traffic is congested during rush hour.)

Is the assumption valid that all of this congestion is caused by the inhabitants? I'd suspect that a lot of the car traffic in Manhattan is created by people from New Jersey, for example, and not by the Manhattanites themselves.
 
Last edited:
all of these feel like super obvious claims to me, the question is, why pick/single out these metrics specifically?

Inner city air quality will improve in the coming decade(s) as more vehicles are going to be electric. The real problem is energy expenditure per capita.
I think another question could be why not single them out? What's the flaw in singling out the aforementioned metrics.
 
I think another question could be why not single them out? What's the flaw in singling out the aforementioned metrics.
choosing metrics reflects your idea of what the problem is

if the authors doesn't make it clear why they single these metrics out, they're not transparent about their stance towards the problem

if you read their study on city traffic (or even the summary) and can't answer the question why they didn't look at time and energy spent on travel per capita, the authors have hidden a bias that should have been made clear if honest intellectual discourse is the goal.
 
Yes, I understand that.

I'm trying to point out that people can use valid data to distract from the issues that actually matter; it's a way to make a bunk argument that can't be debunked by looking at the details (though it's obviously a good idea to do that anyway), but rather by pointing out that it's of the "rearranging deck chairs on the Titanic" variety.

(If I can walk to the nearest grocery store and take the subway to work, I don't care so much that traffic is congested during rush hour.)

Is the assumption valid that all of this congestion is caused by the inhabitants? I'd suspect that a lot of the car traffic in Manhattan is created by people from New Jersey, for example, and not by the Manhattanites themselves.
The main argument that Cox seems to make is that even though the per capita VKT is less in denser areas, the density means more car ownership per area, and more VKT/VMT per area (km2/sq. mi), which cancels the reductions in VKT per capita. This will lead to more congestion as denser areas don't have wide highways to accomodate all that vehicle travel per area.

Now of course this doesn't account for the possibility that drivers in denser areas travel shorter distances, and thus don't spend as much time in congestion as drivers in more sprawled areas.
 
Going back to Litman's data, I will compare it's claimed figures of what the Texas Transportation Institute (TTI) says about 2013 congestion costs:
Screen Shot 2022-09-26 at 1.58.31 pm.png

With the 2013 figures cited in the TTI's 2014 Urban Mobility Report (UMR):
Screen Shot 2022-11-09 at 8.31.40 pm.png
We notice a few discrepancies. The figures cited for delays per commuter are wildly off the mark (e.g. the figure for Atlanta is 39 hours per Litman, while it is 52 hrs per the 2014 UMR). The rankings are rather off (e.g. Boston, New York and Philly rank 8th, 10th and 13th for annual delay hours per Litman, while ranking 5th, 4th and 14th per the 2014 UMR).

Now I don't know where did Litman get those figures from, as his paper only links to the 2000 version of the UMR.
 
I presume it's an inference from the higher vehicle distance driven per area.
yes, but I don't think it's valid?
And why do you think the per area measurement is misleading asides from distance driven?
if you take stats per capita, cities look good
if you take stats per area, cities look bad
if you pick one over the other without a good reason, you're either not thinking, or biased

if you compare a city to a rural area, what you're trying to do is compare what goes on.
but the obvious difference is that there are more people in the city.
there are obviously more murders in NYC than in Podunk, NJ.
but there are also more McDonalds,
because there are more people.
if you want to find out if living in the city causes fast food consumption or murders, you have to account for that population difference,
and per capita numbers do that.

otherwise you're just talking about the obvious.
if you compare NYC with 8 million people to a rural area with 8000 people, NYC is going to have more of everything except crickets and livestock.
so you should
a) either compare NYC with a rural area encompassing 8 million people (like Oklahoma+Kentucky) — care to look up the VKT for these states? or car ownership?
b) or do it per capita

a really interesting statistic, if you're going to compare, is how many hours per week the average person spends on traveling. it doesn't really matter whether it takes you 30 minutes in NYC for a short drive because of rush hour, or 30 minutes in Oklahoma because the store is that far away, does it?
 
a really interesting statistic, if you're going to compare, is how many hours per week the average person spends on traveling. it doesn't really matter whether it takes you 30 minutes in NYC for a short drive because of rush hour, or 30 minutes in Oklahoma because the store is that far away, does it?

If you introduce the concept of a vehicle that uses negligible energy whilst idling, like an EV, then that comparison gets a whole lot more interesting. Time isn't the only cost/waste worth measuring, as energy is also a limited resource. It might be that in the future energy is the more important one to measure.
 
yes, but I don't think it's valid?

if you take stats per capita, cities look good
if you take stats per area, cities look bad
if you pick one over the other without a good reason, you're either not thinking, or biased

if you compare a city to a rural area, what you're trying to do is compare what goes on.
but the obvious difference is that there are more people in the city.
there are obviously more murders in NYC than in Podunk, NJ.
but there are also more McDonalds,
because there are more people.
if you want to find out if living in the city causes fast food consumption or murders, you have to account for that population difference,
and per capita numbers do that.

otherwise you're just talking about the obvious.
if you compare NYC with 8 million people to a rural area with 8000 people, NYC is going to have more of everything except crickets and livestock.
so you should
a) either compare NYC with a rural area encompassing 8 million people (like Oklahoma+Kentucky) — care to look up the VKT for these states? or car ownership?
b) or do it per capita

a really interesting statistic, if you're going to compare, is how many hours per week the average person spends on traveling. it doesn't really matter whether it takes you 30 minutes in NYC for a short drive because of rush hour, or 30 minutes in Oklahoma because the store is that far away, does it?
I think the reason why VMT/area was mentioned is because if you have too much VMT/area, you will place a greater burden on the existing road network than with a lower VMT/area according to Cox.
 
I think the reason why VMT/area was mentioned is because if you have too much VMT/area, you will place a greater burden on the existing road network than with a lower VMT/area according to Cox.
"burden" in the sense of road maintenance costs is majorly caused by heavy vehicles (trucks etc.), and obviously this burden is greater if you have more roads to maintain. Again, here it would be interesting to compare NYC to Oklahoma+Kentucky.

I believe traffic speed is also a factor, a rural highway needs to be in better repair than an urban cul-de-sac, and also degrades faster.

If you have more inhabitants/area, you typically also have more tax income to pay for road maintenance. Which brings us back to why "per capita" expense for transport infrastructure is probably the better basis for comparison.
 
"burden" in the sense of road maintenance costs is majorly caused by heavy vehicles (trucks etc.), and obviously this burden is greater if you have more roads to maintain. Again, here it would be interesting to compare NYC to Oklahoma+Kentucky.

I believe traffic speed is also a factor, a rural highway needs to be in better repair than an urban cul-de-sac, and also degrades faster.

If you have more inhabitants/area, you typically also have more tax income to pay for road maintenance. Which brings us back to why "per capita" expense for transport infrastructure is probably the better basis for comparison.
Sorry. By burden I don't mean road maintenance, I mean traffic volume.
 
Sorry. By burden I don't mean road maintenance, I mean traffic volume.
so you mean road utilisation.
with rail, we were hoping for 100% utilisation; why would high utilisation be a burden when considering roads? They're made to be used.

also, for that to work, you need to correlate VMT/area with road miles/area or even lane miles/area, because cities also have more of that.
 
Back
Top