204 Comments

This is a good post, but I'm not sure bridge failures are the best measure of bridge quality. I'm a bridge engineer and some of the talk I've seen in the industry is more about aging bridges and maintenance costs, basically that inadequate funding for preventative maintenance ends up costing more in the long run. If your maintenance budget is too small then bridges that could have been maintained inexpensively turn into bridges that need expensive repairs ten years from now. The poster child used to be the Greenfield Bridge in Pittsburgh that was crumbling so badly they constructed nets and a smaller bridge underneath just to catch debris. That seems like a good chunk of money that could have been put to better use.

https://en.wikipedia.org/wiki/Greenfield_Bridge

Just want to add a disclaimer that though I'm a bridge engineer I know very little about the national condition of bridges so I'm not trying to speak from authority there, just mentioning what I've seen in seminars and such.

Expand full comment

This does seem like a real issue. I wish we could force state governments to do their accounting with a depreciation paradigm so that deferred upkeep "scores" as costing money over the long-term. Changing the accounting practice might generate a sound case for a short-term funding surge (or else it might not, it's an empirical question) in which case I'd be all for it.

But I think this illustrates the broader point that there's a real need for reforms in the transportation space. Not to say we shouldn't spend any money. But we need to fix our systems to make sure we're in a situation where money will flow to things that are worthwhile.

Expand full comment

I was chair of the capital planning for my town for about a decade. During that time the money available for maintenance never matched the need for maintenance. (In MA a portion of gas tax receipts is allocated for municipal highway repairs.) We did replace one bridge during that period, but two others equally in need of replacement could not be replaced and were just patched and given load restrictions. We prioritized re-surfacing to keep traffic flowing, but the less travelled roads deteriorated every year.

This problem is in part from inadequate funding, and in part from the political reality that big new projects get political benefits. Fixing cracks in the pavement and relocating signs to improve sidewalk snow removal are expenses with no political benefit.

Water and sewer are similarly problematic, and will be getting much worse rapidly. Our leak surveys and historical records showed that most of the water and sewer underground was at or beyond its design life. You don't get much visibility of problems before things break. We could fix the worst leaks. The construction from before the 1920's will break with increasing frequently before we can replace it.

Expand full comment

I wish that could conceivably solve the problem. Why on earth would any politician care about that when simply deferring maintenance makes it somebody else's future problem and they have to find the money now to head it off? It's not an accounting problem. It's a political problem. Here's a dirty secret of the rehab business just for you. You never find out how bad the problems are until you start the work. Which is why such projects almost always go over budget and some times massively so. And projects that frequently go over budget are rarely popular projects for politicians to undertake.

Expand full comment

No politician ever held a ribbon cutting ceremony for fixing a bridge. And most politicians are quite happy that the decay rate is slow enough that they can conveniently kick the problem down the road and make it somebody else's more expensive problem.

Expand full comment

I think the NYC-centrism of the American publishing industry feeds into this a lot, because as you note the BQE really is crumbling, and the general road quality in NYC really is terrible.

But NYC’s roadway maintenance problems are largely administrative and occasionally simply criminal. The various city agencies involved don’t coordinate maintenance timelines so replaced roads sit for weeks without lane markers, milled out roads take weeks to get new asphalt poured, recently resurfaced roads get dug into for utility maintenance, and certain favored neighborhoods get their roads resurfaced regularly while the Bronx rots. And of course sometimes scheduled work just mysteriously doesn’t happen. (Bike path maintenance happens halfway to never, of course.) None of this can be solved by increasing he NY DoT’s resurfacing budget, although increasing the federal DoJ’s appetite for municipal corruption cases frankly might help.

Also I’m not so sanguine that the BQE is going to be rebuilt any time soon. The much more likely scenario to me would be a replay of the fate of the original West Side Highway: a section of it is going to collapse (everybody is worried about the cantilever along the Brooklyn promenade but don’t discount the possibility of one of the concrete supports under the southern stretch through Sunset Park giving way and killing a lot of people in the process) and it’ll sit closed and condemned for a decade or more as lawsuits fly back and forth.

Expand full comment

Good points here.

The thing I would add about NYC-centricity is that the city is *so* large that it's not like you can just dismiss the metro area's problems as insignificant. A problem that's mostly only important in San Antonio doesn't matter. But the New York Combined Statistical Area is 7 percent of the nation's population and over 10 percent of GDP. So it does matter and is worth talking about.

But because the media is based in New York, there's a tendency to implicitly treat it as somehow typical, when it's actually an extreme outlier.

And this creates a situation where even people on the other side of the country underestimate how weird New York is. People in San Francisco talk about "Manhattanization" when their city has about half the population density of Queens!

Expand full comment

My favorite fact to bring up when talking with people complaining about increasing density in SF is that San Francisco is presently less densely populated than notable urban hellscape Somerville, Massachusetts.

Expand full comment

"In the hellscape of Somerville, people live literally on top of each other, three to a building! We can't have that in our beautiful city!" /s

Expand full comment

Frankly the only thing preventing cars in NYC from being even more of a terror are the potholes. Narrowing every street and adding bike lanes and big sidewalks is the only solution I’d support.

The BQE should be removed.

Cars and cities can’t cohabitate.

There solutions are political not financial. NYC has plenty of money for these projects, outside money only ever appears for more highways and wide surface streets. No thanks.

Expand full comment

I suspect you are in the minority on that. People who would oppose your solution are transit riders (buses are big), moving trucks, delivery trucks, emergency service vehicles to name a few. Then there's the problem of moving people with mobility issues and handicaps. Were you planning on making them walk or bike at gunpoint?

Expand full comment

These are not opposing goals. A very good way to make things like emergency vehicles, deliveries and paratransit go faster is to _get the other vehicles out of their way_.

I have a small personal stake in this, as pre-covid I worked (and will hopefully one day soon work again) at an office in Chelsea where a regular feature of the building was listening to 120 dB fire engine sirens directly outside our front window because they were gridlocked on 17th St for 10-15 minutes at a time. This _stopped_ when the 14th Street busway went in and fire engines could cut across 14th without obstruction. My ears appreciated it but I bet the people who had called the FDNY appreciated it a lot more.

Expand full comment

Yeah, basically. Or move.

Expand full comment

I feel like potholes are just as bad if not worse for bikes that for cars.

Expand full comment

Substantially worse, but the bikes to a first approximation don't _cause_ potholes, which is not something you can say about 7000+ pound F-350s.

Expand full comment

No they don't. But the vehicles that do cause potholes are the big heavy ones like buses and delivery trucks. Better eliminate those.

Expand full comment

Yes, I am aware that busses and delivery trucks are heavy. However, they serve a distinctly different purpose in NYC than a raised F-350, and eliminating one out of the three of those things will _reduce_ the amount of damage done to the roads, not to mention the odd pedestrian.

Expand full comment

I recently realized a lot of my personal frustration with America’s highways just stems from I-95 itself. It is maddening that the major artery between the country’s capital and its largest city narrows to two lanes at multiple points. I sit in traffic and think about how China Will Win every time.

But I understand that most people don’t need to regularly make the NYC metro suburb to NYC metro suburb trips that I make, and public transit is just not well designed for my suburb to suburb commute. For most, 95 is an occasional hassle rather than a regular source of frustration.

Still, tractor trailers are just *sitting* in the Bronx for hours every day. Between the local pollution effects and the global CO2 consequences, something needs to be done. And I need to friggin move.

Expand full comment

You are identifying a real problem here, though. New York City is big enough and congested enough that mass transit *should* serve suburb-to-suburb commuting needs.

This is where we could learn a lot from Germany, and seek to knit NJ Transit, MetroNorth, and LIRR together into an affordable, high-frequency, all-electric service comparable to the S-Bahm in Berlin or the RER in Paris.

Expand full comment

as a Westchester resident it's insane to me that one has to go into Manhattan just to get across the county

Expand full comment

It's infuriating, but it's notably hard to solve. Run a line from Peekskill to Katonah and it doesn't help people to get from Yonkers to New Rochelle, or vice versa. You'd end up needing three or four new lines just for the one county.

The better solution is for the three main Metro-North lines to converge on a station somewhere like Wakefield so you can use that as an interchange without needing to go into the city - a track from Yonkers to Wakefield would both create a Yonkers-New Rochelle journey (transfer at Wakefield) and would allow anyone wanting to do suburb-to-suburb on MetroNorth to do so without going into the city. It would work the same way Jamaica does for LIRR and Seacaucus for NJT.

Orbital rail at a long distance from the center (an inner orbital is another question: Staten Island-Brooklyn-Queens-Bronx-Hudson Park-Hoboken-Bayonne-Staten Island could be very useful) doesn't work because going straight through the city is shorter and therefore faster, while for short journeys, the problem is that it won't align with most users actual routes, so people will have to travel in or out to hit the orbital and then the same again to get to where they want to be - and there isn't the density (and therefore demand) to have multiple orbitals. The upshot is that you end up with a lot of people choosing to take the one-transfer route through downtown rather than the two-transfer route on the orbital.

Get in close enough, and there's enough density that it's useful for a lot of people; no-one does more than 45-60 degrees around the circle - it's faster to go in and out again - but there are enough people wanting to go from Queens to the Bronx or from Hoboken to Staten Island that it makes sense.

The other thing about orbital is that it only works if has interchanges with the radials; if you can't transfer then it doesn't work.

Expand full comment

Honestly for cross-suburb transit, light rail or even busses on a dedicated ROW seems like a reasonable idea.

Expand full comment

Depending on the suburb, even just regular buses if you're far enough out that the roads aren't busy.

The usual rule of thumb, though, is that you only have so many ROWs into the city, and suburbs branch from the ends of those ROWs. So you build a station just inside the branching point and then people doing suburb-to-suburb within a relatively narrow angle from the city can change at that branching point, while people wanting to go further (e.g. in a New York case, from Westchester County to Suffolk or Nassau County) do go into the city - remember that congestion on trains doesn't slow the trains down, so it's quicker to go through the city than around it.

For New York, the physical geography is such that it's pretty obvious that you have one of these suburban transfer stations for each of LIRR, NJT and MetroNorth. LIRR has Jamaica; NJT has Seacaucus; MN should have their own. These suburban transfers have the additional benefit that they free up some space on the trains that can be used for urban rail inside the city - people changing at Jamaica free up seats on Jamaica-Penn Station for people who live in Jamaica or along the line at the other intermediate stations.

Expand full comment

also, haven't been in the comments for awhile. what does the blue badge on my profile mean?

Expand full comment

I think we decided that it labels those of us that paid more than the minimum amount for our subscription.

Expand full comment

Oh man, I thought it denoted the superior quality of your discourse! At least I can stop trying to up my game.

Expand full comment

If you hover over the badge, the alt text is "Charismatic Authority" -- I'd been wondering if it indicated you got some minimum number of likes per comment, on average, or something like that.

Expand full comment

Not that it is a perfect project by any means, but DC metro area is getting some exciting suburb-ish to suburb-ish mass transit with the Purple line. I'm just happy about the fact that there will be a line that doesn't just run from some place in MD or VA, to downtown, and then out to some other place in MD or VA. I live in College Park MD and currently taking the metro to silver spring takes 55 min while driving takes 20 if traffic is low. Purple line cuts the metro times down to probably something like 30 min. Even more extreme 1hr20min to New Carrolton on Metro currently with a 20 min drive. Purple line will bring metro time down to something like 20 min.

Expand full comment

The NJT and LIRR tracks even connect under Penn Station--they can through-run those trains now at minimal cost... it's the bureaucratic agencies that need to be figured out (just like your fave--MARC/VRE through-running... dreaming of a Frederick to Frederickburg train...). Is there any hope that a large amount of money for mass transit in the bill would go toward electrification of commuter-rail lines? Seems like a huge value-add to pour money into electrifying these suburban lines in the largest metro areas to ease the transition to RER/S-Bahn style frequent service.

Expand full comment

Most of the big American cities happen to be on a state boundary - NYC's tracks are New Jersey in one direction, and New York in the other; Philadelphia is Pennsylvania one way, New Jersey the other, DC is Maryland one way and Virginia the other. Even when they have track connections in the city, through-running requires two state agencies to co-operate.

One thing that Buttigieg could do is to lean on each of these states to merge the agencies. Split NJT Rail in two, merge half with SEPTA and the other half with the MTA; merge PATH into the NJT-MTA. Merge MARC and VRE with WMATA. Oh, and make the MTA actually merge LIRR and MetroNorth operations.

I still find it amazing that SF managed to end up with two agencies (BART and CalTrain) even without a state boundary. And BART uses a non-standard gauge, so you can't even do through-running.

These mergers don't solve everything (LA has a single rail agency and that doesn't work either), but it does make things like through-running become possible.

Expand full comment

Sadly, the BART design is even worse than just the non-standard gauge - which is indefensible. They also required flat edge rails rather than inwardly tilting rails - so the trains run louder and with more wear and tear so they require custom wheels and some custom ballast equipment that I don't quite even understand. It's a maintenance mess.

https://www.mercurynews.com/2016/03/25/has-barts-cutting-edge-1972-technology-design-come-back-to-haunt-it/

Expand full comment

Chicago also has Metra and the CTA as distinct services despite them almost entirely being in the same state (Metra has a handful of stations in WI/IN). The two don't get along, and usually it's just a hassle, but sometimes it leads to real waste.

The CTA wants to spend a few billion dollars to extend its Red Line (main north-south line) further south, which sounds OK until you realize that Metra already has commuter rail stations with stops a few *blocks* apart in that set of neighborhoods already, which serve a line going downtown as well.

It would clearly be a better solution to spend a few billion upgrading Metra infrastructure, tracks, etc. to get faster headways and turn it into an "urban rail"/RER-style service, but the CTA doesn't want that kind of encroachment on its turf.

Expand full comment

If ONLY the sfbay area had merely two transit agencies! The last time I looked, the 511.org page listed well north of TWENTY of them between Napa and San Jose.

Even if you want to restrict it only to train operators (which would be weird since you're excluding AC Transit at that point), you've also got the VTA light rail in Santa Clara and Amtrak's Capital Corridor line which is a de facto commuter rail between Oakland and Sacramento.

Expand full comment

The contrast between this and Transport for London or RATP is really striking.

There are a few turf wars between TfL and some of the urban rail operators, which are regulated by the (national) Department for Transport. And RATP and SNCF hated each other (RATP built the Métro with right-hand running while SNCF is left-hand running so that they could never be physically connected), though RER (which is shared between the two) has broken down those barriers in the last few decades.

But a big metro-wide transit authority with control over buses and subway and trams and light rail and some urban rail and then a turf war with a national longer-distance rail agency (note that the private UK rail operators are effectively contractors to the national transportation department) is the normal approach right across Europe. Some outer suburbs may have local buses under local control - but this would mean, from the SF perspective, something like Sacramento having its own bus system, not Oakland.

Expand full comment

I95 literally destroyed Philadelphia and turned its most valuable land (the riverfront) into a toxic waste site.

A journey that used to take weeks taking an extra hour is not a big deal. Turning our most productive places into brain destroying lead and carcinogen spewing dystopias has made us all much poorer.

Expand full comment

Amen on all counts. The Cross-Bronx gets all the attention thanks to NYC's cultural centrality and of course Robert Caro's reporting, but for my money what I-95 did to Philly was every bit as bad and then Philly covered itself in shame by doubling down on it _over and over again_, first by creating the Vine Street Expressway in the 90s and now with the "95revive" project (which at least is finally proposing capping some of it in Center City) and a million other projects in between.

Expand full comment

Should just tear the entire thing out between the Betsy Ross and Walt Whitman 295 in NJ can be the new 95.

Expand full comment

As is often the case in these debates, we are still paying for the sins of Robert Moses and his contemporaries in the postwar era. The fact that I-95 goes _through the middle_ of both NYC and Philadelphia is just utter madness and can’t easily be fixed.

Give me a magic wand and/or a trillion dollars, and I’d tear out the Cross Bronx and re-route I-95 across the Tappan Zee and GSP, bypassing NYC entirely. And then toll the length of it with toll rates rising with the AQI.

Expand full comment

Sorry if this is a stupid question, but would it actually cost money to remove highways from densely populated neighborhoods of NYC? Surely the development rights to the land are worth billions.

Expand full comment

Oh yes indeed. You could probably solve NYC's housing crisis in a single stroke by tearing out the Deegan and the FDR and zoning the underlying land at FAR 20.

The problem is that of course there's a deadlock of at least three independent bureaucracies involved here and none of them have any incentive to do this and in fact most of them are designed to make sure it never happens: it took the second-worst earthquake of the 20th century to get the Embarcadero Freeway torn down in San Francisco and even that remained a right-of-way for cars afterward. NYC is quietly more NIMBY than San Francisco: the next time (it's not an if but a when) we lose a major highway due to structural damage, you can expect that they will propose surface streets and parks, but housing only over the dead bodies of the entire city government.

Expand full comment

Matt and other YIMBY folks haven't talked much about replacing roads with buildings. It seems as if it might be both harder and easier than the policy changes YIMBYs do recommend.

Expand full comment

In general that's right, but Alon Levy does have a suggestion to do this for the BQE: https://pedestrianobservations.com/2019/04/03/shut-down-the-brooklyn-queens-expressway/

Expand full comment

Can't zone for FAR 20 residential in NYC without state law change :(

Expand full comment

😭😭😭

Expand full comment

That is a great suggestion!

Expand full comment

There is a road that does this: I-287. It branches off from I-95 at the NY/CT border, goes over the Tappan Zee bridge, and then hooks down into NJ and dumps into 95 again south of Princeton. A bit west of the GSP but otherwise exactly as you suggest, and it is heavily used by trucks looking to bypass NYC.

Expand full comment

Yes. I, a resident of NYC, suggesting a diversion of I-95 over the Tappan Zee, am aware that that road is currently designated as I-287. 🤣

Expand full comment

It’s insane that the highest toll route in the US is in sparsely populated PA and not frickin NYC.

Expand full comment

Right??? Frankly every highway and bridge in NYC should be _aggressively_ tolled, and the tolls should rise during rush hour and by AQI. Driving into lower Manhattan at 8am on a day when the AQI is over 75 should cost more than a taxi from JFK to midtown.

Expand full comment

Don't forget that I-95 itself wasn't supposed to funnel into the NJ Turnpike. Up until a few years ago, you couldn't even drive all the way from DC to NYC on I-95!

https://www.theatlantic.com/technology/archive/2018/01/after-61-years-americas-busiest-highway-is-almost-complete/550982/

Unfortunately one of the side effects of the lack of building the Somerset Freeway was the increased truck traffic on US 202 and 206 through that part of NJ. Not sure if that's a point in favor or against the induced demand argument...

Expand full comment

Yeah the Cross Bronx expressway should be ripped out.

Expand full comment

I wonder what is in those trailers? Food maybe? I foresee problem feeding people without trucks.

Expand full comment

It's one thing for trucks bound from the port at Elizabeth to the Bronx to be stuck waiting in the Bronx. But it's another for trucks bound from Philadelphia to a Walmart in western Massachusetts to be stuck waiting in the Bronx. I believe much of the original interstate plan had beltways around the cities specifically to avoid that, but the later expansions made the interstates run right through the cities, and many of the cities have grown up to include their beltways, so that now the long-distance traffic causes congestion within the cities.

Expand full comment

Why yes. They need a new by-pass or two. Now try building them.

Expand full comment

"My daily commute is down to my basement where Slow Boring Headquarters is located."

A vertical commute. Do you use an elevator, helicopter, or an Elon Musk rocket-powered hyperpod?

Expand full comment

"Quick, Robin! To the Bad Take! uhh... Bat Cave!"

Expand full comment

Hey Matt, you probably already know this, but the NYT Morning Newsletter today quoted your Substack from yesterday pretty substantially.

Expand full comment

I saw that too - congrats, Matt!

Expand full comment

How did years of living in DC get you to mellow out on sprawl? For me, dating somebody with parents were way out in the MD burbs and seeing those big houses that were relatively new but already beginning to fall apart; listening to a boss complain about their 90 minute commute and then complain about people riding their bikes to work; hearing friends who moved to the burbs talk about how much they hated coming into the city to do stuff; hearing those same friends brag that their rent is $200 less than it was in the city, but also complain about the cost of car insurance; etc etc...has only made my rage grow. Maybe it shifted from a cultural rage to the same kind of rage that makes me mute "dog" on Twitter--a more precise, needling rage.

Aside from the rage point, when you're on your way up to Maine you should try to drive though Albany if you ever want to see a beautiful city truly destroyed by Robert Moses-think.

Expand full comment

What I've come to appreciate about sprawl, is that while I don't think the DC situation is optimal I do think that the greater sprawlness of DC compared to New York or San Francisco has generated a better affordability situation for working class people.

Expand full comment

and compared to sprawling places, like, Dallas or Atlanta, DC is completely unaffordable for working class people

Expand full comment

Yea but Fredrick MD is affordable and part of the commute pattern.

Expand full comment

The Frederick-DC commute is *rough* and not recommended for someone who wants to keep their sanity.

Expand full comment

On the other hand, Philadelphia is less sprawly than DC (you can build taller here and the suburbs don't go as far out--there isn't a substantial community that commutes from Reading or Atlantic City to Philadelphia), but is substantially more affordable. Maybe not as much as Texas, but it shows that the correlation isn't one-to-one.

Expand full comment

yep, the two big variables are supply and demand of housing. You don't need sprawl to get sufficient supply, but you limit infill and limit sprall and expect affordability

Infill (Philly, or Tokyo if we're being extreme) > Sprawl (DFW, ATL, HOU) > Coastal NIMBYism (SF, LA, DC, BOS, etc.)

Expand full comment

One thing that shocked me from One Billion Americans is that Philadelphia has lost a third of its population from peak. That should make things much more affordable in and of itself.

Expand full comment

Household sizes everywhere have fallen by quite a bit - I believe I've seen factoids talking about DC and parts of Los Angeles, where the number of housing units has gone up, and vacancy rate has gone down, but population has also gone down, because five person families have been replaced by couples and singles.

Expand full comment

That's interesting and I hadn't thought of it before. Do you know a good way to check the data on household growth instead of population growth?

Expand full comment

I feel like defending sprawl is some next level contrarian take given that your readership is filled with YIMBYs. Sprawl can only be tied to affordability if you consider the alternative to sprawl 'build nothing'. But there's a better alternative and in practice 'affordability via sprawl' guides society away from the better alternative.

And I mean, the term itself is pejorative, which also makes it a weird thing to defend. Even the people who actually kinda do love sprawl don't say they love sprawl.

Expand full comment

I mean, I feel like a lot of people still like suburbs despite the discourse.

Expand full comment

THIS! So much discussion about how cities don't allow infill, which is bad. But so little recognition that many people like living in suburbs. They don't like traffic, but they like having space that living in the city doesn't allow.

Expand full comment

Speaking as someone who will happily cop to being one of the more aggressive-to-insane YIMBYs here...

Suburbs are fine. If you like suburbs, you should definitely move there! Many of my best friends, etc etc etc.

What drives me 'round the bend is that we effectively _force_ people to move to the suburbs by making development in our thriving cities impossible. Then we force people _further_ out by making further infill development in the inner-ring suburbs illegal. And we spend billions to trillions of dollars subsidizing this madness in a dozen different ways.

Expand full comment

I had to move 60 miles outside Philadelphia to find a real estate market that wasn't bananas. I can't even imagine the considerations that make people find these truly expensive cities seem reasonable.

Expand full comment

2-3 hours each way in the bay area and you do things like start work at 6 am you avoid rush hour. and yes that means you wake up at 3 to be on the road at 4 am.

Expand full comment

As a political matter in many communities the alternative sprawl is build nothing or almost nothing.

Expand full comment

This can be a sad political reality, but it doesn't mean we have to 'appreciate' sprawl as anything beyond 'not the actual worst thing that we could be doing for poor people'.

Expand full comment

Well it depends I look at accepting Sprawl vs nothing the way I look at accepting my least favorite Democrat over a Republican I am going to do it give the guy money phone bank etc. It kind of depends on how much you believe electric cars will mitigate the climate costs. If you think not at all then you might be ok with making everyone live with their parents and have lower overall economic growth. Fredrick MD is affordable and long commutable to DC.

Expand full comment

That feels a bit like comparing apples to oranges to me in that the demand in places like NYC and SF seems much greater than in DC. But I appreciate the point that if increasing density is not politically viable then sprawl is the main way to increase housing stock. It also helps meet the innate desire, lurking within most people, to hate their neighbors and own a lawn mower.

Expand full comment

Honest question: how is the "induced demand" argument suggesting that building more roads will make traffic worse functionally different than the NIMBY argument that building more market rate housing will make housing issues worse?

Expand full comment

Well the big conceptual difference is that it's a lot more plausible that a household could be induced to switch from driving 5 miles per day to driving 10 miles per day than that they can be induced form occupying one home to occupying two homes.

Expand full comment

Induced demand makes me angry because it is seen as a bad thing to allow people to go to soccer practice during rush hour.

Expand full comment

Well I think it is seen as a bad thing to the extent it motivates such bad rush hours to exist in the first place, and in the soccer practice example, motivates whether or not to carpool or play for a more nearby team or what time practice is scheduled for.

Expand full comment

Expanding capacity for peak driving is bad yes, we should use pricing instead.

Expand full comment

OTOH, it's a lot more plausible a new home could induce someone to move to a city than that a new highway could

Expand full comment

A new highway could easily induce someone to move to a farther out suburb - which is just one of the ways that it induces someone to switch from driving 5 miles to driving 10 miles.

Expand full comment

This makes sense if these are new roads all together, but if they're additional lanes to already existing roads, this makes no sense.

-let's say I drive X miles per day

-a new lane is added to a highway, so traffic is lessened

-because there's less traffic, I drive X+Y miles per day

-everyone else starts to do this, so traffic gets bad again

-I go back to driving X miles per day

Eventually we'll reach an equillibrium that's no worse than the status quo, right? (Because this rests on the assumption that this incremental driving is due to less congestion.)

Expand full comment

I think the idea is that we have some model where people tend to gradually increase the distance they drive, until the marginal time cost of the additional increase reaches their limit. If that limit is at the same level of congestion, then by the time that limit is reached, the total amount of vehicle miles traveled is much higher, because you've reached the same congestion level on a larger amount of roads.

The question is whether this has happened because everyone has found better matches of their ideal coffeeshop, workplace, park, etc. that are all a bit farther from home than the second-best one they had been going to before; or whether this has happened because everyone has relocated their home to a location that was briefly better under the temporary low-congestion environment, but is now worse in every way because the congestion is just as high as it had been.

Expand full comment

I never understood the “induced demand” point. If people drive more when you build more roads, maybe traffic didn’t get any better, but apparently there were other benefits to people driving more that were worth it (or they wouldn’t do it). So building the roads was still worth it.

Expand full comment

I think the point is that largely the people that demand new lanes are the people that are already driving because they want less traffic. Adding lanes does not solve those people's problems, and in fact just adds to the people unhappy with the traffic. The new drivers induced by the new lanes are absolutely choosing to do it, but frankly they were probably fine without the new lanes, so while you've maybe slightly improved the well-being of the new drivers, the old drivers are equally or even more unhappy (because driving on 4 congested lanes is probably slightly more unpleasant than driving on 3 congested lanes). All the while, you've increased the number of drivers causing negative externalities that are totally unaccounted for (air pollution, CO2 emissions, congestion, and worse walk/bike-ability). Now if all those externalities were accounted for and drivers still wanted more lanes enough to pay for them themselves, then sure go for it. But until then, adding lanes doesn't solve the problem people advocating for them want solved, has opportunity costs compared to other projects, and increases the number of people causing the negative externalities of driving, in exchange for maybe making a relatively small amount of new drivers slightly happier.

Expand full comment

You do a good job enumerating the pros and cons, but I don't see why we should expect the new driver benefits to be so minimal. I'd guess they're often worth it.

But sure, if your actual concern is there should be less traffic then charging congestion fees is probably more effective than building new roads.

Expand full comment

Or of course both! I’m sure this will get most pro and anti-car people upset...

Expand full comment

I always thought induced demand, when used in reference to expandable major multi-lane expressways, is basically about equilibrium. When there's more demand than capacity on major roads, people use less efficient local streets or reduce discretionary trips, and then if more lanes are added to the most efficient major expressways, the pent-up demand uses the new capacity until it again reaches equilibrium and pushes traffic onto other routes, etc.

Expand full comment

This sort of discounts all the negative externalities of freeway development. Decreased land available for high value housing, health outcomes near mass amounts of car exhaust. And, invariably, no city has found the point where lanes have solved the problems for more than a year or two.

As a city dweller, I'm sort of tired of people ripping up my city to make it more coinvent for exurb folks to drive into my home.

Expand full comment

This sort of discounts all the negative externalities of freeway development. Decreased land available for high value housing, health outcomes near mass amounts of car exhaust. And, invariably, no city has found the point where lanes have solved the problems for more than a year or two.

As a city dweller, I'm sort of tired of people ripping up my city to make it more coinvent for exurb folks to drive into my home.

Expand full comment

It’s often significantly more challenging to change the job or sell the house once your commute has gotten bad.

LA et. al. have continued to solve this by sprawling and building more lanes. It’s seems to be a solution that simply doesn’t scale.

Expand full comment

Yes the capacity increase helped someone. If you want to sell more roads by making it easier to live a driving oriented lifestyle be my guest, but you're never going to solve congestion, you'll just make it worse.

Notably nobody ever sells roadway expansions via capacity increases, because other drivers don't want to spend money on more new drivers and to worsen congestion!

Expand full comment

Yes, I think this is largely right. Though if some of those shifts were shifts to new locations that *looked* better under the temporary low-congestion conditions, but turned out not to be better once everyone is stuck driving farther, the new equilibrium could end up being worse for everyone. A lot depends on what the empirical facts are of which new VMT are created.

Expand full comment

the concept is that they drive more *because* traffic gets better. So if so many people now drive more and traffic gets worse, then they would drive less.

If this refers to brand new roads that go to brand new places, I can see how that would be different. But going from two lanes to three lanes (or three to four, or whatever) can't theoretically make traffic worse based on that simple assumption, right?

Expand full comment

I think it can make things worse. This would be "Braess's Paradox". I think a lot of people claim that this happens far more than it actually does, but an example of how it can work is that if you add more lanes to a freeway, but don't demolish any city buildings to increase the space on local streets or create more parking, then some additional suburbanites might drive into town for morning errands and afternoon soccer practice - although the highway is equally congested to what it was before (or maybe even somewhat less), the city streets are more congested, and the people who were driving entirely locally now deal with worse traffic, despite having no benefit from the freeway lanes.

Expand full comment

I've been arguing about this "Induced Demand" thing in the comments here at SB for a while, and I don't want to repeat myself *too* much, very briefly, Induced Demand is actually:

a) At least 3 separate things not one thing

b) Sometimes greatly overstated

c) Not always a bad thing when it does happen

d) Unfairly applied only as an argument against road building / improvement

Expand full comment

As someone who does believe it is at least sometimes an important under-considered problem, I do agree with you about all of these points.

Expand full comment

One "cost" of using a road is congestion. So more lanes make driving "cheaper" since we don't charge by the mile. The public pays, but the individual uses it. New market rate housing charges rent.

Expand full comment

I'm confused.

More lanes means its less costly (in terms of congestion) to drive. So people may, on the margin, drive more. But the idea that this will mean that traffic will get worse doesn't make sense given that this incremental driving was only due to congestion being less bad. If congestion gets bad again, people stop doing that incremental driving.

Not sure how the concept of who pays for things fits in here.

Expand full comment

When you pay a cost in money, at least *someone* benefits from your spending. When you pay a cost in time, *no one* benefits from your spending. Money costs are socially neutral, except for distributional impacts, while time costs are socially negative.

Expand full comment

I think the argument is that if you go from two lanes to three and you add 50% more cars, then you haven't improved anything.

Now, the argument that the extra people who are now using that highway and weren't before are benefitting (because otherwise they wouldn't use the highway) is one that is well-taken.

But the argument mostly used is "this road is congested and it takes me too long to drive along it, widening it would make it less congested so I'd get there sooner" - and induced demand generally means that they don't actually get there any sooner, because the limiting factor on the traffic was the congestion.

At peak hours the demand for the road if it was uncongested may well be five or more times the current capacity, so the only limiting factor is people who don't drive (or don't drive at the time, ie timeshift their journey) because of the congestion. Increase capacity, and you just soak up that suppressed demand.

That isn't necessarily a bad thing - the same applies to public transit; if the subway is full and crowded, people are avoiding it and taking the bus or walking or cycling or travelling at a less busy time; add more cars to the subway, or run the trains more frequently and those people now switch to the subway that they prefer to the other modes.

The problem is that it creates more VMT and we know that VMT have lots of externalities.

Expand full comment

You are correct. If someone is saying that adding a third lane will increase usage by more than 50%, then I don't agree with that either.

Payment matters in that anything where the user pays zero will tend to result in queues. Queues for auto travel are called traffic jams.

Expand full comment

So when traffic is really bad you will only drive it if you have to for work. You will not plan your soccer practice during rush hour you will swipe left on the cutie if she is too far away. That is what is meant by induced demand. You might think it is ok to drive 30 miles for a first date at 5pm on a weekday. To me that sounds like a major quality of life improvement. The alternative is to have congestion pricing so the poor still cannot enjoy life from 3pm-7pm but the rich can.

Expand full comment

I think you want to separate the problem of poverty from congestion pricing - we need some way to say "hey, there's actually a pretty big cost to you using this road at this time, you should pay it or figure out another road/time to use". Currently we basically ask people to pay for this with time in most places by sitting in traffic; congestion pricing seems like a more effective way of getting people to understand the tradeoff and look for alternatives. The problem of poverty should be addressed by tax/wage/etc. policy, not by transportation policy.

Expand full comment

Yea I guess I just just don’t mind over building highways so people can do fun stuff at 5 pm.

Expand full comment

People are coming up with overly complicated answers to this. The reason is actually simple: using roads is almost free, but rent costs money.

It's the same reason you can't open an ice cream stand in the middle of nowhere and all of the sudden have created demand for more ice cream than you can make. But if you gave it away for free, people would consume more ice cream.

Expand full comment

Yeah I know demand curves slope downwards.

Are you saying that if a place had congestion pricing, so driving did cost money, would adding highway lanes suddenly not induce demand?

Expand full comment

Yes!

Expand full comment

seems like that would depend on the level of congestion pricing and how elastic that is.

Expand full comment

Building more roads results in more driving, which is widely agreed to be a bad thing.

Building more home results in more people living in home (or bigger homes, or homes closer to the places they want to live), which is widely agreed to be a good thing.

Expand full comment

People that like the lifestyle of driving and commuting exist. How can you shop at Costco and take $280 in groceries home on the train?

Expand full comment

Sure. But everyone agrees that the actual process of driving to costco is just a cost you pay - if it was a 5 minute drive instead of 20, the experience would be even better.

Expand full comment

People's lifestyle habits end up filling up the new road capacity. This is addressed very well in "The Myth of Travel Time Savings"

https://www.tandfonline.com/doi/abs/10.1080/01441640701642348?journalCode=ttrv20

Basically you add a new highway lane. Things briefly get better, suddenly people start making trips to use the extra capacity, and reorient their lives around it, congestion goes back to the same previous level and everyone spends one average 30 minutes commuting one way.

Expand full comment

I think, empirically, the former appears to have some (limited) truth to it, while the latter does not.

Expand full comment

I'd be interested to see that research. Not to be too cynical, but it seems from my vantage that being YIMBY (which is good) but also being against adding lanes to existing roads is more of an urbanist anti-car aesthetic thing more than anything.

Expand full comment

Another theory explanation is that there is a price mechanism for allocating scarce homes but no price mechanism (except congestion pricing, which does solve congestion) for driving.

Expand full comment

not sure how that functionally makes a difference. Incurring a cost of sitting in traffic and having to pay higher rents are costs either way that we'd like to avoid

Expand full comment

When a cost is incurred by sitting in traffic, that is human life just thrown away. When a cost is incurred by paying higher rents (or congestion prices), the money is just transferred to someone who then gets some value from it. So from a social point of view, monetary costs are net neutral (except in the distributional effects) while time costs are net negative.

Expand full comment

The throughput on roads drops dramatically once you put too many cars on it. So using wait times (congestion) as the way to clear the market dramatically hurts the total output of travel. Adding more roads doesn't improve the issue once a city becomes big enough because it still falls into the slow traffic jam equilibrium.

Expand full comment

how though? Is it because congestion follows an exponential curve and adding lanes to highway are linear improvements or something?

Expand full comment

Three comments.

1) While not NECESSARILY a problem, I’m suspicious of a bill that seems to imply that part of the benefits is to create jobs. Jobs are a COST in a cost benefit calculation. And if the spending pattern increases demand for high-skilled jobs more than low-skilled jobs, that is rather anti-redistributive.

2) We ought not be too concerned about the amounts per category. If there is a good selection procedure, we will not invest “too much” on any category. And if the procedure is bad, even less money will not be well spent.

3) I am concerned that investment in outmoded federal infrastructure (e.g. IRS computer and auditing systems) and maintenance does not get a category.

Expand full comment

Jobs are a cost only from the perspective of a private enterprise. From the point of view of a redistributing state, jobs are a benefit unless you are crowding out private investment. As a state, you have to pay jobless people. If you give them jobs, you still pay them but get their labor in return.

Expand full comment

No, jobs are a cost even to government. And even to a redistributing state. When a state employs people to do useless jobs the nation ends up poorer than it would be if the state employed people to do productive jobs.

Expand full comment

Good post. Let me again emphasize the importance of port improvements. The US is an import dependent country. Most of those imports arrived by sea in shipping containers. The global logistics system is under tremendous stress right now. The Suez Canal blockage was another Brock in the Wall. Another stress point is LA/Long Beach where dozens of ships are idling offshore. Increasing the throughput of American ports is imperative.

Expand full comment

What's the issue with the LA ports?

Expand full comment

It is the most important container port in the US. It is currently suffering from amazing levels of congestion. There are currently dozens of ships waiting to unload their cargos. These ad their burden of engine emissions to the air of the LA Basin. However, the drama is not over when the containers are offloaded. Few of these containers are destined for Southern California. Instead they must be shifted to the

Railheads for cross-country transit. Ideally, this transfer would be ship directly to railcar but that's not the way it works there. The railyards are miles away. So the containers are loaded onto diesel powered trucks. They are then driven through LA's notoriously congested throughways, adding yet more hydrocarbon emissions to the basin. Your former colleague, Dave Roberts, had suggested replacing diesel drayage trucks with electric. That would reduce emissions and increase energy efficiency. However, it would not eliminate the congestion issue. Relocating the rails would require eminent domain seizure of some very expensive property with the usual NIMBY response. However, that would be the best solution. More dockside cranes, a major piece of capital investment, would also be helpful.

Expand full comment

Because there is such a back up in Long Beach, many container ships have been re-routed north, and are now idling in SF Bay as well.

Expand full comment

Not just currently; when I lived in LA 20 years ago congestion was a huge issue that was in the news all the time and a thing (local) politicians talked about.

The idling container ships were causing so much particulate emissions that there was an unmistakable rise in asthma and all kinds of childhood and developmental disorders centered on San Pedro. It became a case study in particulate air pollution and all the horrible things that correlate with it.

Expand full comment

I did not know that the rails do not go up to dockside.

That seems jaw-droppingly stupid.

I can see why the NIMBYs might not want multiple freight rail lines in their backyards. So why not just dig out the LA basin all the way to the railyards? Everyone loves water-front property.

Expand full comment

At one point there was serious talk about burying the 710 because adding lanes was rendered nigh impossible by NIMBYism. I think Matt has written about this before, but you can stall large projects indefinitely in California with red tape, environmental impact surveys, etc., not to mention all the local objections. Extending the rail lines to the port was (and probably still is) a perennial traffic-reduction promise of politicians that never came to fruition.

Expand full comment

Didn't construction of the grade-separated Alameda Corridor help with this in Los Angeles--connecting the ports by rail to the railyards on the Eastside?

Expand full comment

I confess I don't get this post:

America's commute times are short (because it relies more on cars than all the other countries?) but what that says about the conditions of the roads, I have no idea.

People don't like paying higher taxes and that means they're fine with the roads (or maybe they want to get things for free, or want to eliminate the "waste, fraud and abuse"?)

The average roughness of roads has decreased, but left out is that (mostly state and local) spending on operations and maintenance has increased (but maybe the federal government should cover more of that with zero real interest loans and let states and localities divert money to other crucial things?)

The ASCE raised our grade from D- to D so everything is fine now. Not that I believe them when they say we're in terrible shape so why should I believe them when they say things are better?

The money from increased taxes going to highways etc would be better used for, e.g., child poverty (probably true) -- but it's 5% of the total bill, so perhaps isn't the best target for reallocation.

Well, I guess every post can't be a winner.

Also, this:

>>but the biggest issue is that even as Los Angeles built out their transit infrastructure, they didn’t change land use policy to encourage dense development near the stations.<<

Well, there's some truth here, although LA is taking steps to increase density around transit stations (such as the TOC program). Much more can and should be done. But that's not the biggest issue. Far bigger is that LA is simply really really big geographically. The city itself is 503 square miles, compared to 234 for Chicago. The Los Angeles-Long Beach-Anaheim metropolitan statistical area is 4850 square miles. You'd have to build a *lot* of track to cover extensive areas like these. Unfortunately, most of LA Metro doesn't connect people where they live, and where they work, and other places they want to go. That's one of the reasons why ridership is so low.

Expand full comment

"The ASCE raised our grade from D- to D so everything is fine now. Not that I believe them when they say we're in terrible shape so why should I believe them when they say things are better?"

They are subject matter experts who overvalue what they have to offer but are likely to be directionally correct. It's like the public health recommendations for COVID: almost always too strict but skilled enough in interpreting the data well to move in the right direction

Expand full comment

I don't think it matters what *fraction* of the city LA Metro connects - all that matters is what *number* of people and potential destinations it connects. 93 miles *should* get you a lot of connections, and it can if they increase density around those stations. (That density will also enable non-transit connections between those nodes and their immediate surroundings.)

Expand full comment

I think the fraction does matter. Yes, there are places where major loci of employment are really close to rail stations -- USC is right on the line, and most of Santa Monica is close to the E-Line terminus. Downtown LA has a bunch of stations that serve people who work there. But the LA area is *so* spread out that the chances of building lines that link places where lots of people work and lots of people live are pretty low, until the system is much more widely built out. As it stands now, major places of employment are simply not well-served by rail transit.

Sony Culver City is a 1.3 mile walk from the closest station. UCLA is 2.2 miles away. CBS Television City is 4 miles and Disney/ABC in Burbank is 5 miles.

Some of this will improve over time (e.g., entertainment firms will be occupying the new offices adjacent to the Culver City station, there will be a Westwood/Wilshire station (so UCLA will only be a half mile from the rail line)), but building a robust network sited near large numbers of jobs will take a long time, as will building increased density of housing near stations.

Expand full comment

The LA vs. Chicago thing is also vehicle ownership rates.

I can't find another more recent study I saw on this but the link below estimates that in 2016:

27.5% of Chicago households were without a vehicle vs.

12.2% for LA, 10.1% for Long Beach, just 5% for Orange County

This huge disparity creates a way higher base rider load for the CTA.

https://www.governing.com/archive/car-ownership-numbers-of-vehicles-by-city-map.html

Expand full comment

I'm worried that we're once again focused on the "last war" and not true upcoming realities. The budget for repairing and upgrading our electrical infrastructure is a step in the right direction, but with impending self-driving car technology coming (never as soon as Elon says, but still...), shouldn't we be planning for THAT? The cost of Uber/Lyft/TeslaRide will then be competitive with public transport costs (since no need to pay a driver), and why would I spend two hours on public transit when I can spend half that in a self-driven car for my commute for around the same money? Roads will become more important again, public transport less so.

Expand full comment

The first step to ignoring a problem is pretending it doesn't exist. You have done a fine job here with some of the goofiest charts I have ever seen. Oddly this does not mean I disagree with the funding levels advocated in the Biden plan for fixing 'crumbling' infrastructure in the way of roads and bridges. It just means you don't have a clue what you are talking about when it comes to what crumbling might mean. What, pray tell, is a bridge failure? It falls down? Or it is deemed unsafe to use? If the former the only acceptable standard is zero. The basis of that particular chart is ludicrous. Failures per million kilometers of road is a stupid way to measure anything. The US has a lot more roads, particularly highways, and a lot more bridges or overpasses per kilometers of road than designs used in other jurisdictions. Which means you have more in absolute terms than anyone else. Seriously, who thinks the Russians are stunningly good at building and maintaining this sort of infrastructure?

The fact that the highway and road system is well designed and suits most people's needs is a good thing. They are. But they must be maintained. I do not think you understand the significance of the road roughness chart either. The fact that it is trending downward is a good thing. But that is entirely due to the recognition that rough roads carry immense costs in transportation efficiency, vehicle maintenance but most importantly lives. Rough roads cost lives. I note that this piece seems to conflate simply fixing decaying infrastructure and other improvements to such infrastructure. Don't do that. No one is talking about adding lanes to your crumbling New York highway or any other such thing. We are talking fixing what exists. Which is perfectly possible, necessary unless you want it falling on your head, but expensive. The longer you delay it the more expensive it gets.

I am glad you are taking the whole induced demand argument with a grain of salt. You should, it is a dumb argument. If you build more road capacity you get more traffic? No kidding. Exactly the same thing happens to every transportation system. I don't know who thinks that if you increase density and max out your subway capacity in place like New York you won't need to figuratively build more lanes. And man is that expensive and time consuming. So I have sad news.

To those people who admire the big comfy seats on Amtrak trains I have similar bad news. If it carried the traffic to make it economically viable those big comfy seats would go away fast.

Expand full comment

My understanding is that it's best to think about Amtrak as two entirely separate train systems that are frankenstein-attached to each other: the northeast corridor, which would be nicely profitable if run on its own and which indeed probably should be both shrinking seat size and decreasing train headways in search of more profits, and... "all the rest of it", which is basically a rolling vacation system for train hobbyists both as passengers and as staff, and which should probably just be quietly shut down.

Expand full comment

I think the San Jose/Oakland/Sacramento line and the Los Angeles/San Diego line are close to the margin of profitability as well.

Expand full comment

Entirely plausible!

Expand full comment

I've never understood the argument for VMT tax, at least so long as fossil-fuel cars remain the norm. A gas tax effectively taxes the miles you drive divided by the fuel efficiency of your vehicle, which seems optimal, encouraging people to drive less and / or use a more fuel efficient vehicle.

I understand the issue in the US is that the current gas tax comes up a little short in terms of highway funding. In many countries the situation is the opposite, and what motorists pay into the system in terms of Vehicle Tax + Fuel Duty + VAT (usually paid *on top* of fuel duty on fuel, plus on other car related items) + Tolls + Congestion Charges often greatly exceeds what is actually spent on the road network.

Expand full comment

There are two reasons to tax driving. one is the Pigovian tax on the burning of fossil fuels. The second is to pay for the use of roads. The gas tax use to do both jobs pretty well but with greater fuel economy and electrification, the two purposes have uncoupled, You can make the high fossil fuel users subsidize the others beyond the Pigovian tax (which should be much higher than it is at present) but it's probably not good economics. The Pigovian tax should pay for general decarbonization or other social benefits, the VMT directly for roads

Expand full comment

This is a fair point. But like I say most European countries seem to have no problem politically with gas and other taxes that go beyond what is even required to maintain highways (not a fan of this just how it is).

Expand full comment

Matt--you say the US is a laggard in employment to population ratio, but the study you link to doesn't say this. The study just shows that other countries are catching up to the US--it shows rates of change, and some cross-sectional levels, but not overall levels.

As of 2013, the US has the 18th highest employment to population ratio in the world per wikipedia. It's slightly below the G7 average, but is above the median (the average has a positive skew from very high rates in a couple countries).

Expand full comment

I was going to ask whether lower employment-to-population ratio was the explanation for lower commute-time-per-population, but if the employment-to-population ratio isn't actually lower, then that goes away.

Expand full comment

I've noticed that roads are a good bit smoother in Germany. I don't know how much more that costs, but it is nice.

Expand full comment

Like you, I kept contrasting my experience of daily driving on I-95 and I-93 around Boston to what I saw when I lived in Germany and Austria in 2002-2008.

Rusting bridges and potholes here, freshly painted bridges and smooth roads there.

I am curious why the posted data here is so at odds with the personal experiences of many.

Expand full comment

Another view, from Washington State.

State's bridges, roads crumbling faster than they can be fixed

The Ship Canal Bridge on I-5 carries a quarter-million people every day between North Seattle and downtown. It's wearing out, and ignoring its severe deterioration will have “dire consequences,” a state report warns. In fact, it adds, Washington would need to spend an estimated $14.8 billion over the coming decade to achieve “minimally acceptable condition” for roads, ferries and bridges. Will this be the year lawmakers stop procrastinating? Here’s what they have in the works. (Photo: Steve Ringman / The Seattle Times)

Expand full comment