Tag Archives: freeways

The Hollywood Strangler – Part 1

The Ventura Freeway, running from Woodland Hills to Pasadena, is the Valley’s main – only, really, since the 118 is so far north – east-west freeway. In typical LA fashion, much of the Valley is car-oriented despite the paucity of freeway capacity: just one east-west freeway for nearly 2 million people. The freeway is the famous 101 from Woodland Hills to the Hollywood Split, and the much less famous 134 from the Hollywood Split to Pasadena.

The Ventura Freeway is newer than some of LA’s first freeways, like the 110, the 10, and the Hollywood Freeway portion of the 101, so it doesn’t suffer from the problem of short interchange spacing. However, it’s old enough to have underpowered freeway-to-freeway interchanges at the 405, the Hollywood Split, and the 5. Let’s take a closer look at these three interchanges, as usual with an eye on rationalizing the freeway facility and improving the local streets in the vicinity. For this post, we’ll look at the Hollywood Split, saving the 405 and the 5 for another day.

Before diving in, let’s think about what makes a good freeway-to-freeway interchange. First, these interchanges have larger ramp volumes than typical interchanges, which makes the weaving conflicts worse. That suggests an increased need to avoid conflicting local ramps (a local on-ramp right before the freeway off-ramp, or a local off-ramp right after the freeway on-ramp). Second, these interchanges take up more space than typical interchanges. Therefore, in order to maintain functionality for local traffic distribution, it often makes sense to have an interchange serving local traffic integrated somehow. This can take the form of half diamonds or a full diamond interchanged arranged so that its ramps don’t cause any weaving conflicts. I like to call this an “inside interchange”. Here’s an example: the 134 and the 2, with inside diamonds on the 134 at Harvey and the 2 at Holly.

Now, we don’t want to go blowing massive holes in North Hollywood to drop in 65 mph ramps and create interchanges like they have in undeveloped parts of Fontana. However, these principles can still help us figure out what will work for these cases. We’ll do our best to keep improvements within available right-of-way, compromising on speed where needed. Alright, on we go.

The Hollywood Split is the somewhat confusing junction of the 101, the 170, and 134 in North Hollywood. The 101 enters from the southeast on the Hollywood Freeway and departs to the west on the Ventura Freeway. The leg to the northwest is the 170, and the leg to the east is the 134, both of which end at the interchange.


Despite being signed as the mainline freeway, the 101 exits on the right and merges on the right going east/south, and exits on the right and merges on the left going north/west. This is contrary to modern design standards, which require that the mainline freeway stay left, with exits and entrances to the right.

On the other hand, the interchange also reflects downtown-oriented design, with the movements to/from downtown emphasized at the expense of other movements. This reflects the thinking of the era, that people would drive to a downtown central business district (CBD) in the morning and out to suburbs in the afternoon. With LA’s polycentric development, downtown is not as dominant as it is in many cities. The Ventura Freeway provides important east-west connectivity to outlying CBDs in Sherman Oaks, Burbank, Glendale, and Pasadena. Thus, if one considers the Ventura Freeway the mainline – an argument for which there is a good case, as we shall see – there are only two through lanes, which is also substandard for this location. The Ventura Freeway through movements stay left at the splits entering the interchange, but merge to the right departing the interchange. This means that depending on which movements dominate, we might make different decisions about what to consider the mainline freeway, and which ramps to reconfigure.

Looking at the freeway-freeway ramp layout, we can see that the northwest and southeast quadrant ramps are missing (north to/from west, south to/from east). Again this was fairly typical for that era of freeway design, but leaving out ramps is frowned up these days as it is confusing for motors and shunts high speed traffic onto local streets. The southeast quadrant ramps are more consequential, because they would connect major nodes (Hollywood to Burbank & Glendale), while the missing northwest quadrant ramps would connect lower density areas. However, it’s obvious that there’s very little right-of-way available for the missing ramps, and it might be hard for the southeast quadrant ramps to compete with Barham and Forest Lawn, which make a relatively uncongested shortcut serving these movements.

Lastly, looking at the local street ramps, things are actually in pretty good shape. The 170 has a half-diamond to the north, and the 101 has a half diamond to the west. The 134 has a half diamond to the east, though there’s a little friction between the Vineland on-ramp and Cahuenga off-ramp going east. One notable gap is that there’s no on-ramp to the Ventura Freeway west between Pass Ave and Tujunga Ave or Moorpark St, almost 2 miles, so that might be something to try to fix. Going south on the 101, the Vineland Ave off-ramp is uncomfortably close to the 101/170 merge, something we looked at fixing in a post that feels like it was written century ago.

Now, to look at the deficiencies of the interchange, it’s helpful to look at a stylized diagram showing the number of lanes and traffic volumes. Note that the ramps are drawn as simply as possible, ignoring loops and bridges, to make things easier to look at.


(Note: traffic and ramp volumes from Caltrans. Asterisk indicates volumes I increased by 10,000 to get consistent results.)

It’s readily apparent that the central deficiency of the Hollywood Split is that there are only 2 through lanes on the Ventura Freeway (the 134 west to the 101 north and the 101 south to the 134 east). The traffic volumes are more or less evenly split between the 134 and the 101: 68,000 from the 101 south to the 134 east and 62,700 remaining on the 101 south; 63,000 from the 134 west to the 101 north and 64,400 continuing from the 101 north. In order to handle their traffic volumes, these 2-lane ramps would have to flow full for 16 hours a day. Meanwhile, the ramp volumes between the 134 and the 170 are about half of what the movements are in the other directions.

For the purposes of this post, let’s do an updated stylized diagram showing a simple solution for this bottleneck, taking the easiest approach. Let’s bump up the ramps between the 101 and the 134 to three lanes each way. Going west on the 134, that means we just get rid of the lane drop and extend the third lane through the interchange, which would be a basic bridge widening project. This would leave us with 6 lanes going north on the 101 instead of 5 lanes; the sixth lane could be dropped at the next off-ramp, Laurel Canyon Blvd.

Coming the other direction, let’s pick up the southbound onramp from Laurel Canyon Blvd to the 101 south as a sixth lane. We can then split and have 3 lanes go to the 101 south and 3 lanes to the 134 east. We eliminate the 134 east offramp to Riverside Drive, which has relatively low volume, and let that traffic be picked up by the Tujunga Ave and Cahuenga off-ramps. That clears space for the bridge widening for the third lane to the 134 east. With 2 lanes merging in from the 170 south, we have 5 lanes on the 134 east. Rather than drop the right lane and add the HOV lane on the left, let’s just turn the left lane into a carpool lane and force ramp traffic from the 170 over, since it has a lower volume anyway.


Note that we also cleaned up the 101 north to the 170 north transition, making it 3 lanes instead of 4 lanes and eliminating the need for a lane drop on the ramp from the 134 west to the 170 north. In this instance I left the right-side lane drop and added the HOV lane on the left, but maybe the opposite approach would work.

In a future post, I’ll lay this out on an aerial, and take a deeper look at some of the other possibilities for improvements discussed above.

Let’s Go Glendale!

Having bid a fond “see ya around” to Palms, we turn our eyes to observing Glendale and getting to know this part of the LA region better. An outcome of LA’s legendary traffic and underpowered transit is that it can be punishing to try to experience parts of the region far from where you live. The Valley isn’t that far from the Westside, but the 405 makes it seem far. That problem certainly applies to travel between Palms (the Westside) and the Burbank-Glendale-Pasadena area, which stands out even among the many difficult trips in the region.

For readers outside Los Angeles and not familiar with its confusing municipal boundaries, we should perhaps first explain where Glendale is located. Glendale is a separate incorporated city, not part of the City of Los Angeles. Downtown Glendale is about 8-9 miles due north of downtown Los Angeles, though the city’s northern reaches extend over 15 miles from downtown LA. Glendale borders the cities of Burbank, Pasadena, and La Canada-Flintridge, along with an unincorporated neighborhood of LA County known as La Crescenta-Montrose. Glendale also shares two borders with the City of LA – Sunland-Tujunga to the northwest, and Atwater Village, Glassell Park, and Eagle Rock to the south. Lastly, Glendale’s northern limits extend up to the Angeles National Forest in the San Gabriel Mountains. The Verdugo Mountains separate downtown and the southern part of the city from the northern part, located in the Crescenta Valley, a narrow valley between the Verdugos and the San Gabriels.

An unconventional way to define Glendale might be as the valley of the Verdugo Wash. This is a short tributary of the LA River that joins the river near where it takes a sharp right turn from running west to east through the SF Valley and heads south to downtown LA. Like the LA River, it is fully contained in a concrete flood control channel. The Verdugo Wash runs east to the north of downtown Glendale, then gradually turns northeast, north, and northwest as it wraps around the mountains of the same name into the Crescenta Valley. Everything south of the 134 – all of downtown Glendale and many residential areas – actually drains away from the Verdugo Wash, but topography makes one suspect that this area is sort of an alluvial fan deposited by the stream. There might be potential for improvements to the Verdugo Wash like those proposed for the LA River.


The primary freeways serving Glendale are the 5, the 2, and the 134, which roughly form an upside down triangle around downtown Glendale. Despite serving Glendale, these portions of the 5 and the 2 are almost entirely in Los Angeles. North of the 134, the 2 continues north through the more mountainous portions of the city, ending at the 210, which serves the Crescenta Valley.

Traffic on the 5 is perhaps not quite as bad as the 10 and the 405 on the Westside, but it’s bad enough. Since the 5 runs the full length of the Golden State, it seems to have a larger volume of background traffic, and a notably higher amount of truck traffic – even if your carpool, like mine, leaves at 5 am. Truck traffic is probably increased by the gap in the 710, which eliminates a potential route around downtown LA between the ports and destinations to the north.

The 134, together with the 101 in the Valley and the 210 east of Pasadena, forms a long, continuous east-west freeway stretching from Ventura to San Bernardino, another heavily used corridor in a region with no shortage of well-used freeways. While the 101 and the 134 in the Valley and the 210 east of Pasadena get heavily congested during peak periods, the 134 between Glendale and Pasadena seems to escape the worst traffic. Astute eyes will note that the short Colorado Street freeway, connecting the 5 to San Fernando Rd and Colorado St in Glendale, looks like an abandoned attempt at routing the 134 through the heart of downtown Glendale. In fact, Caltrans’ small white bridge identifying signs still mark these structures as being located on the 134, so there’s potentially a companion post to Walk Eagle Rock’s post on the 134 being rerouted to avoid downtown Eagle Rock. The selected route for the 134 is not only better for downtown Glendale, but much better for a freeway network than the puzzling location of the Colorado St freeway’s end at Griffith Park.

The 2 is perhaps best known for the portion of the freeway that wasn’t built – the portion from the existing end in Echo Park to the west, through Hollywood, Beverly Hills, and Century City to the Westside. This leaves the extant part in Northeast LA and Glendale as one of the more lightly used parts of LA’s network, though congestion on connecting freeways like the 5 can turn parts of it into a giant queue. It’s also the reason it’s hard to get to the Westside from Glendale in the absence of good transit options.


Ok, enough about freeways, let’s get on to the things that will really interest readers here: transit. At first glance, your LA Metro map makes things look pretty good.


However, what we have here is a classic case of wide coverage with relatively poor frequency. Here’s a look at some important routes serving Glendale.


Routes 90 & 91 serve Glendale Ave, which runs to the east side of downtown and the Crescenta Valley. Route 92 serves Brand Blvd, which is Glendale’s main commercial street. Route 94 & Rapid 794 form a very long route from downtown LA to the independent City of San Fernando, near the northern end of the eponymous Valley. This serves only the western edges of Glendale, but it’s the closest route to me. Finally, Routes 180 & 181, & Rapid 780, serve east-west travel between Pasadena, Glendale, & Hollywood.

Evening and late night headways fall off pretty quickly, making it tough to depend on these routes if you want to do anything other than work your 8 to 5. The two Rapid routes, 780 & 794, don’t run at all late nights or on weekends. Rapid 780 runs with good peak frequencies, and because it’s through-routed as the Rapid for both Routes 180/181 and Route 217 (Fairfax), it sort of functions as the transit route doing what the 2 freeway was supposed to do. (Don’t bother with Route 201, which only runs hourly.) Therefore, when Rapid 780 isn’t running, riders face an additional transfer between Routes 180/191 and Route 217. On top of that; there are the usual reliability issues; on a recent weekday morning my Next Trip app promised 794 service in 42 minutes and 57 minutes. You can sort of see why the BRU complains about this when rail riders get 10-12 minute headways all day, every day.

On the rail side, Metrolink offers a Glendale station at the very southern edge of the city, adjacent to Atwater Village. Frequencies during peak periods are pretty good – there are 30 trains per day – but service ends early, going to hourly or worse at about 6:30pm and ending altogether at 9:30pm. The worst feature of Metrolink is the absurd pricing; a one way ticket from Glendale to Downtown LA is $5.50 to travel a distance of 6 miles, a distance you can double or triple on Metro rail lines for $1.75.

The upside of all of this is that there’s a lot of low-hanging fruit for transit improvements in the area – things that don’t involve, say, building an expensive underperforming light rail line to bridge the gap in the 710 freeway.

As a first take, transit improvements should include improving frequency and spans of service. Options to improve reliability, such as bus lanes and signal priority, should also be explored. On the rail side, Measure R2 plans should explore upgrading these Metrolink Lines to rapid transit frequencies, though that should probably be contingent on upzoning some of the land near the rail corridors.

Development Patterns

Speaking of development, let’s talk a little bit about the built environment in Glendale. As mentioned before, Brand Blvd serves as the heart of downtown, with Glendale’s small skyscraper district (five buildings of 20+ stories, six more of 15-19 stories, almost all outcomes of the late 80s boom) centered on Brand and the 134 freeway. Downtown Glendale has been undergoing a residential and mixed-use mini-boom, with Americana at Brand being the best known development. Since there are many projects in progress or recently completed, it’s probably worth doing two separate posts, one on the commercial projects built in the 1980s and early 1990s, one on the ongoing residential projects. Some people deride Glendale as boring, but having spent a couple evenings on Brand Blvd, I’m willing to say they either don’t know what they’re talking about or are using “boring” as code for “full of retail establishments but not the kind that I like”.

Outside of downtown, there are residential neighborhoods that are actually somewhat similar to, well, to Palms. The residential density of the Census tract I moved to is only a little bit lower than that of the tract I moved from. The biggest difference is that the percentage of single-family residences (SFRs) in my new neighborhood is higher than in Palms, where you might miss the remaining SFRs if you didn’t know where to look. The apartment building stock in Glendale also appears to be newer, with few dingbats and more apartments dating to the 1980s boom, something supported by a casual look at Property Shark. Nevertheless, I’ve done the math, and my apartment building’s 9 units on a 50’x150’ lot are exactly classic R3 dingbat density. When I walk around, though, none of the remaining SFRs are being replaced by apartments, and at first glance the zoning appears to make even existing apartments non-conforming. I’m sure there’s a fascinating story behind that, one we’ll no doubt have to explore in more detail in a future post. . .

The 101 – Four-Level to Hollywood Split

‘Bout time for a freeway post, no? Prerequisite: principles for urban freeway improvements.

When it opened in 1954, the 101 was a revolution in freeway engineering, incorporating many lessons engineers had learned on earlier efforts. Gone were the tight corners of the 110, and interchanges now had acceleration/deceleration lanes. Breakdown lanes also appeared.

However, the 101 also features the typical shortcoming of 50s-era freeways: interchanges are spaced too close together. Maybe engineers expected lower speeds or smaller volumes, or maybe they just didn’t understand the impact of weaving. Either way, with distant rumblings about the need to widen the 101 – an expensive and disruptive endeavor for sure – it’s time to offer an alternative that helps improves traffic and coincidentally improves the city too. (Or, look at the other way around if you want.)

One good aspect of 1950s freeway engineering is that it usually didn’t blow enormous holes in the street grid to shave a few seconds off motorist delay at offramps and onramps. Instead, ramps are cleverly feathered into the street grid. Consider a typical interchange on the 101 (left) compared to the 105 (right).


So the biggest thing to fix on the 101 is interchange spacing. As a general rule, urban freeway interchanges should be spaced like rapid transit stops: about every mile. Ramps spaced less than a mile apart are counterproductive unless you go to the expense of braiding them, which costs a lot of money and disrupts the city – the opposite of what we want to do here. Don’t take my word for it, ask the folks who wrote the book on freeways. A quick look at the 101 shows that they apparently stuffed in as many ramps as they possibly could, regardless of whether the local street warranted it.


First step, figure out the interchanges. We’ll revisit the design of each in more detail later. Courtesy of our friends at Caltrans, here are ramp volumes for the 101. I used all 2006 data, because this was the most complete set. Numbers marked with a star are 2007 data. The changes since 2006-2007 have been trivial, and reflect a time of very high unemployment anyway. I’ve arranged the data so that complimentary moves are next to each other (e.g. northbound off and southbound on).


Shooting for an interchange every mile, and trying to can the low volume ramps first, here’s my rough layout. Mileposts start from the freeway center of the universe in East LA.

  • Four-Level Interchange (MP 3.1)
  • Alvarado (MP 4.4)
  • Silver Lake (MP 5.3, NB off & SB on only)
  • Vermont (MP 5.9)
  • Santa Monica/Western (MP 7.3)
  • Hollywood (MP 8.0)
  • Franklin/Vine (MP 8.6, NB on & SF off only)
  • Cahuenga (MP 9.0, SB off & SB on only)
  • Highland (MP 9.5)
  • Barham/Universal Studios (MP 11.1)
  • Campo de Cahuenga (MP 12.1)
  • Hollywood Split (MP 13.0)

Ten miles, eleven interchanges.

Note: in all the graphics, green is freeway replaced with undeveloped land. Orange represents land to be developed with new buildings.

Four-Level Interchange

I’m still leaving this one alone for a future post. For now, let’s say that obviously, we’re going to have an interchange with the 110.

Alvarado St

The ramps to/from Echo Park Av (northbound) and Union Av (southbound) are too close to both the 110 ramps and the Alvarado ramps, and they dump freeway traffic onto local streets. The NB off/SB on pair has higher volume than Alvarado, but that may be due to people using those ramps to avoid traffic on the 101 as soon (or for as long) as possible. No one using those ramps would have to go more than 0.7 miles out of their way to get to Alvarado.

Therefore, the Echo Park and Union ramps should be eliminated. The street grid would be reconnected and some property would be redeveloped. No changes are needed at Alvarado, which is a tight diamond.


Silver Lake Blvd

The ramps to/from Rampart Blvd and Benton Way are too close to both the Alvarado ramps and the Silver Lake ramps, and they too dump freeway traffic onto local streets. Closing these low volume ramps would be only a minor inconvenience to those drivers, who would have to travel a maximum of 0.6 miles to get to Alvarado. These ramps should be eliminated.

Meanwhile, the Silver Lake ramps are heavily directionally biased – that is, volumes to and from the south are much higher than volumes to and from the north. (See the ramp volume table; NB off/SB on are much larger than SB off/NB on.) The SB off and NB on ramps are also very close to the Vermont ramps, making for a punishing weave section on the 101. While it’s usually bad practice to not offer all ramp movements at every interchange, in this case an exception is warranted. The SB off and NB on ramps at Silver Lake should be eliminated. No one would have to drive more than about a mile out of their way due to this change.

Again, this would allow the street grid to be reconnected and some property to be redeveloped. No changes are needed at Silver Lake, which is a tight diamond.


Vermont Av

Vermont has by far the highest volume NB off/SB on ramp pair, and the SB off/NB on ramp pair volume is large as well. This interchange is definitely staying.

That creates an interesting question regarding the ramps at Melrose and Normandie. The SB off/NB on ramp pair volume isn’t that bad, but the NB off/SB on pair is high – in fact, the second highest such pair in the corridor. Eliminating those ramps is a tall order, but if they’re not eliminated, they probably need to be braided – the opposite of what we’re trying to accomplish here.

As luck would have it, an accident of history created an opportunity to solve this problem. The median of the 101 is very wide, up to 160’, between Virgil and Melrose. Students of LA freeway history will recall that this is because it was planned to be the location of a freeway interchange between the 101 and the 2, which was to be extended from Echo Park west to Beverly Hills and Santa Monica.

That freeway is never getting built, and the wide median is currently occupied by a self-storage facility. Time to put it to better use.

Between Virgil and Vermont, the southbound side of the 101 would be shifted north. There’s already an empty bay in the middle of the Vermont Av bridge that was reserved for the 2 freeway, so there’s no need to do any work on it. From Vermont to Melrose, both sides of the 101 would be shifted towards the middle, eliminating the wide median. New frontage roads would be constructed from Vermont to Melrose, accommodating the traffic currently using the NB off/SB on pair at Melrose/Normandie.

Finally, the Vermont ramps would be reconfigured to square up the intersections, eliminate traffic lights, and free up land for development. In particular, note the relocation of the SB onramp to line up with Rosewood Av and the new frontage road. This allows Oakwood St to be reconnected and for new development on the north side of Oakwood to screen the freeway.


Santa Monica Blvd/Western Av and Hollywood Blvd

Things are a little bit simpler at Santa Monica, Western, and Hollywood. There’s no need to modify those interchanges. The partial interchange at Sunset/Wilton has high volumes for the NB off/SB on pair, but it’s just too close to the Santa Monica/Western ramps. I’d expect most of that traffic to shift to the Hollywood interchange, which currently has smaller volumes. The SB offramp to Sunset isn’t a huge issue because it’s braided, but its volume is pretty low. That ramp and the SB onramp from Hollywood would be modified into a frontage road and onramp, reconnecting a north-south street in the adjoining neighborhood.


The changes on the southbound side aren’t that big, but eliminating the loop on the NB offramp frees up a lot of land for development and is a huge improvement to the pedestrian environment at Sunset and Wilton.

Franklin/Vine and Cahuenga

The ramps to/from Gower St and Argyle Av are all low volume ramps. They can be eliminated without much issue. The SB offramps to Cahuenga and Franklin/Vine are both high volume ramps, as is the NB onramp from Franklin/Argyle, so they remain. The SB onramp from Cahuenga should remain just because there needs to be a SB onramp somewhere in the area. The NB offramp to Cahuenga should be eliminated, since it is a low volume ramp and creates a murderous weave with the high volume NB onramp from Franklin/Argyle.

Another option would be to replace the braided SB ramps, on from Cahuenga and off to Franklin/Vine, with a frontage road and SB onramp. This would eliminate a couple bridges over city streets, and free up land on the north side of Franklin for development.



The SB off/NB on ramp pair at Highland is the highest volume ramp pair on the whole corridor, and it’s in a very logical and undisruptive location. That pair is definitely staying. The NB offramp to Highland needs to stay, just to have a ramp in the area, since the Gower and Cahuenga offramps would be eliminated. It’s closer than ideal to the Franklin/Argyle onramp, but so it goes.

The SB onramp from Highland should remain but be closed except when events are ending at the Hollywood Bowl. It’s much too close to the Cahuenga offramp for everyday operations, but it makes sense to keep it for Hollywood Bowl events so that the large bursts of traffic from those events don’t have to go through city streets to get to the freeway.


The NB off/SB on pair at Barham has fairly high volumes. The NB on ramp at Cahuenga near Pilgrimage Bridge also has high volumes and should remain. The NB offramp to Universal Studios is lower volume, but can remain because it’s not taking up useful real estate or causing any major weave problems. The NB onramp at Universal Studios should also remain.

For reasons unknown to me, the original NB off/SB on pair at Barham was abandoned long ago – all the way back in 1957, according to California Highways. If it can be reactivated, the original NB offramp would increase weaving distance between the Cahuenga onramp and the Barham offramp. However, that might encourage drivers entering northbound at Highland to try to cut across to Barham, which is a major weave and therefore undesirable. Instead, perhaps the Cahuenga onramp to northbound could be relocated south to increase the weaving distance between it and Barham.

Campo de Cahuenga

The Lankershim NB off/SB on ramp pair has larger volumes than Campo de Cahuenga, but it’s too close to the Barham ramps. In addition, eliminating the ramps at Lankershim would improve the pedestrian environment. Pedestrian volumes should be lower on Campo de Cahuenga because there’s nothing there. A SB offramp should be constructed at Campo de Cahuenga, and all movements provided for the NB ramps. The Campo de Cahuenga interchange would replace both the Lankershim interchange and the Vineland interchange, which is too close to the Hollywood Split.


This frees up land for redevelopment on Vineland where the SB loop ramps would be eliminated.


Hollywood Split

The Hollywood Split, with its missing ramps and left-hand exits and entrances, is a project unto itself. I’m going to leave it alone for the time being.

Implementation Timeline

Some of these improvements require building new ramps, which would require more traffic studies and engineering design than easier locations where ramps would just be eliminated. There’s no reason to hold up the easy projects to wait for the more complicated ones to be ready for construction; the easy ones could be completed in less than a year. A few improvements could be accomplished in the midterm, e.g. closing braided ramps with bridges that would need to be demolished. I’d break things down as follows.

Short-term improvements (<1 year):

  • Close Echo Park Av & Union Av ramps
  • Close Rampard Blvd & Benton Way ramps
  • Close Silver Lake Blvd SB off/NB on ramp pair
  • Close Normandie Av/Melrose Blvd SB off/NB on ramp pair
  • Close Sunset Blvd/Wilton Pl NB off/SB on ramp pair
  • Close Gower St/Argyle Av NB off/SB on ramp pair
  • Close Cahuenga Blvd NB offramp
  • Close Highland Av SB onramp
  • Close Vineland Av NB offramp

Mid-term improvements (1-3 years):

  • Close and demolish Sunset Blvd SB offramp
  • Close and demolish Gower St SB offramp

Long-term improvements (3+ years):

  • Close Normandie Av/Melrose Blvd NB off/SB on ramp pair, realign the 101 mainline lanes, and build frontage roads between Vermont Av & Melrose Blvd
  • Close and demolish Franklin Av/Vine St SB offramp, construct SB frontage road between Cahuenga Blvd & Franklin Av/Vine St, and reconstruct Cahuenga Blvd SB onramp
  • Complete tight diamond interchange at Campo de Cahuenga, close Vineland  St SB off/SB on ramps, close Lankershim Blvd ramps

Staging the implementation allows low-cost improvements to be constructed first. In fact, almost no upfront expenditure would be necessary for the short-term improvements if the land was sold for development, since the developers would assume the cost of removing the existing ramps. This would create some immediate benefits while long-term improvements were studied in more detail.


I know a lot of people really, really don’t like urban freeways. But just saying “tear them down” is a throwaway response that ignores the related issues like land use controls and the role of transportation in cities and economies. It also misses out on opportunities to improve the city that can be made quickly and low cost. The 101, as LA’s first real freeway, is a great place to start, and a successful project here could lead to more short-term improvements across the city.

Why Are Highways Numbered to Satisfy Road Geeks?

In the US, we generally sign three different kinds of numbered highways: interstates, US routes, and state routes. In some states, there are also signed numbered county routes, but we’ll ignore them for now.

The numbering schemes for interstates and US routes (generally) follow some basic rules:

  • Two-digit routes are main routes (one-digit routes are understood to be included here).
  • Odd numbered two-digit routes go north-south.
  • Even numbered two-digit routes go east-west.
  • Interstate numbers increase south to north and west to east.
  • US route numbers increase north to south and east to west.
  • Three-digit routes are of the form “xPP”, and are supposed to connect to their “parent” two-digit route “PP” somehow, e.g. US-119 is supposed to connect to US-19.
  • Furthermore, the “x” for three-digit interstate routes is supposed to be an odd number if the route is a spur that connects to the parent only at one end, and even if the route is a loop that connects to the parents at both ends. For example, the I-405 freeway connects to the I-5 freeway at both ends, while the I-710 freeway only connects to the I-10 freeway at one end. . . sort of.
  • Two-digit numbers are almost never used in more than one place. Three-digit numbers are reused.

That’s all well and great. . . if you’re trying to make road geeks happy. In fact, road geeks get angry about routes that do not follow the numbering system; see the fury drawn by I-99 and the US-4xx routes if you don’t believe me.

Problem is, satisfying road geeks is a pretty strange goal for your highway numbering scheme. The purpose of building highways, as I understand it, is to facilitate the movement of people and goods. Highway numbering should promote that purpose. No one is out there driving around aimlessly looking for the 5, and, upon finding the 405, feeling relief that the 405 must take them to the 5. Actually, thanks to some rule-bending, a person depending on the numbering scheme would find themselves hopelessly lost in many places; consider, for example, that the 278 in New York doesn’t connect to the 78 anywhere. . . it doesn’t even come close.

Worse than that, in some parts of the country, the insistence on trying to follow the scheme makes things more confusing. The two prime examples of this have to be the Bay Area, which has only one two-digit interstate (the 80) and is therefore drowning in a sea of x80’s (the 280, the 380, the now-defunct 480, the 580, the 680, the 780, the 880, and the 980) and the Hampton, VA area (the 64, the 164, the 264, the 464, the 564, and the 664). When you roll up the 101 into San Jose, do you benefit from having to choose between the 280, the 680, and the 880? Of course not. Some states have implicitly admitted this; see, for example, that Maryland doesn’t sign the 595, realizing that the DC/Baltimore area, and the Northeast Corridor in general, already has plenty of x95’s.

So what would a logical highway numbering system look like? Well, the first thing it should do is communicate the quality of the road to you using the route shield. We have that a little bit with interstates – when you see the red and blue shield, you know you’re getting a freeway. But, in the same way that not all rectangles are squares, not all freeways are interstates.


So, any limited access facility should get the interstate shield. In the sign display shown above, the 5, the 101, and the 60 are all the same quality of road. They are all completely controlled-access freeways. The 39, which is implied to be just as good as the 60 by its route shield, is just a regular surface road.


Ahh, much better! You’ll note that I didn’t bother to change the 101 and the 60 to follow the interstate numbering scheme, even though there are numbers available (the 60 could be the 410, the 610, or the 810, and the 101 could be the 705). That’s because the interstate numbering scheme is pointless. There’s no logical reason that numbers have to increase south to north and west to east, and there’s no logical reason to not reuse numbers if they’re separated by a large distance. The 57 freeway should just become the I-57 freeway. No one is going to get confused and think they are in Cairo, IL instead of Orange County.

That takes care of freeways, which should all get the interstate shield, and state routes, which should be normal arterial roads. What about US routes? They should be used for major cross country routes that aren’t freeways. The current scheme basically allows the use of the US route shield if the route crosses a state line, which leads to the ridiculous scenario where the 199, at 80 miles long and serving the middle of nowhere in California and Oregon, is worthy of a US route shield, while the 99, which is over 400 miles long, is mostly freeway, and serves Chico, Sacramento, Stockton, Modesto, Fresno, Manteca, Bakersfield, and LA (almost) isn’t. The 199 probably does not deserve to be a US route. The 99 does.

Really, a lot of the 99 should be an interstate. The parts that are a freeway should use the interstate shield. The parts that aren’t should use the US shield. There’s no reason that one consistently numbered route has to use the same shield the whole way. Using different shields would help drivers understand that, say, the Angeles Crest Highway and the Glendale Freeway aren’t the same kind of road, even though they’re both part of the 2.


Nothing wrong with that signage.

There is another legitimate reason to do this: it takes godlike transportation power away from highway geeks (AASHTO) and faceless, unaccountable hacks (Congress). The interstate shield, because it symbolizes quality roadway transportation, carries a lot of weight, and states are willing to spend money to get it. That allows AASHTO and Congress to have discriminatory power to force some states to spend money for no good reason. Anyone want to make the case that it is really critical that Caltrans upgrade a few ramps on the 210 in San Bernardino before they’re allowed to throw up the interstate shield? Especially in light of comparison to laughably deficient legacy freeways like the 278 in New York and the 70 in Pennsylvania?

Granted, this isn’t a huge issue. We’ve got better things to worry about and spend money on. But as signage is replaced over time, we should move to this kind of system.

All sign images generated with Kurumi’s sign maker app.

Should We Worry About Highway Subsidies?

I touched on this issue way back when I wrote about the gas tax, but I’d like to expand the thought.

One of the most common criticisms of auto infrastructure from transit and smart growth activists is that drivers don’t pay the full cost of roads – the gas tax and other associated fees have not been increased enough to keep pace with spending new construction and backlogged maintenance. Much of the money spent on highways comes from property taxes, which you pay regardless of if or how much you drive. Counter to this, you have folks like Randal O’Toole, who note that no transit agency in the country covers even its operating costs with fare revenues, let alone capital costs. Transit agencies don’t pay property taxes, but they run buses over roads paid for by those taxes. In addition, the federal government and many states have dedicated part of their gas tax revenues to transit, meaning that drivers subsidize transit.

Still Not User Fees

As I said in my post on the gas tax, I don’t see how transit activists can win under the “user fee” framework. Some, like Cap’n Transit, claim that transit would make money if drivers were forced to pay the full cost of driving. However, given typical farebox recovery ratios on US transit systems (about 25%-50%), I don’t see how that could happen. Assuming farebox recovery is currently 50%, an agency would have to either double the number of people on each vehicle, double the amount of money extracted from each rider, cut unit operating costs in half, or some combination of the three. (Note that just doubling ridership doesn’t cut it if you have to run additional vehicles, since that costs money.)

Realistically, it seems to me that a scenario in which a transit agency has 100% farebox recovery is a scenario in which low ridership routes are eliminated, low ridership stops are eliminated, off-peak service is reduced, and peak service fares are higher. Now, maybe you’re fine with that scenario, but you should back up, read your Jarrett Walker, and ask yourself what you’re actually trying to do with your transit service. Are you ok balancing the transit agency’s books by raising fares on people too poor to afford cars? Are you ok with stranding people who live on low volume routes? Are you ok telling your city’s late-night crowd to suck it up and pay for a cab?

Generally, though, the agency is being asked to provide some minimal level of service to all parts of the region, for some minimum span of service, regardless of profitability. In that context, it’s not consistent to expect the agency to be profitable. There are also many benefits that accrue to society as a whole that the agency can’t capture – for example, if someone chooses to ride transit instead of driving, there are benefits to air quality from less congestion. In that sense, we aren’t “subsidizing” transit, we’re making an investment in the public domain that ought to produce future public benefits exceeding the cost.

And here’s the thing: many of the same arguments apply to roads.

For example, implementing tolls or increasing the gas tax is only progressive at the crudest level of analysis. In general, transit riders are poorer than drivers, but there is huge variability within drivers. Within the driving population, these taxes might be regressive, since wealthier drivers can afford to live closer to work. Like low volume transit routes, it is expensive per capita to provide arterial roadways to rural areas, but we’ve decided that in our society everyone deserves some base services. We also expect roads to produce benefits to society that aren’t directly captured by the government agency in charge of roads – for example, when rubber-tire internal-combustion trucks became available, there was a large reduction in the amount of horse poop lying in city streets. (The memories have faded, so we don’t often think of the horse poop benefits of trucks nowadays.)

Public Services Framework

In fact, both roads and transit could be considered public services like police and public schools, and we certainly don’t expect the police department or elementary schools to fund themselves entirely from user fees.

In that case, why charge drivers anything for road use (or why charge patrons anything to ride transit)? There are two reasons to charge for public goods:

  • Negative externalities (in this case, mostly air pollution and GHG emissions)
  • Overuse (in this case, congestion)

With this framework, the gas tax serves both purposes: it imposes a base usage fee that discourages people from driving for no reason, and it taxes people in proportion to the amount of pollution they create. The gas tax should probably be increased nationally because of the high costs of air pollution and GHG emissions. Some states or metro areas might consider a further increase as a base congestion charge. Managed toll lanes, like exist on the 91 and the 110, should be implemented on a larger scale to help deal with congestion during peak periods.

Another nice feature of this framework is that it’s perfectly logical to charge drivers more than it costs to maintain the road if demand is very high. The surplus can be used to fund other parts of the transportation system. For example, New York charges very high tolls on the Hudson River bridges and dedicates the surplus to transit operations. It’s also reasonable under this system to charge wealthy Acela patrons more than it costs to run those trains, and subsidize other services.

It always seems like a pretty cynical argument to me when I hear transit activists argue that “drivers should pay the full cost of roads”. Under a counterfactual where highway user fees generated more than enough money to cover maintenance of existing roads, would they be arguing that the rest of the fees should be used on roadway expansion capital projects? Of course not. Taking roads and transit to be public services results in a more consistent argument.

What About Overbuilding?

Part of the argument is that if drivers had to pay the full cost of roads, we’d build less roads. True, and valid if your preexisting goal is building less roads. By the same token, if transit riders had to pay the full cost of transit, we’d be building fewer trophy streetcars and suburban LRT lines.

Overinvestment and misallocation of resources is a classic problem of public services. Cities with useless streetcars are no different than rural towns whose police equip themselves with tanks or cities that say they’re going to supply every student with an iPad. In other words, there is no substitute for good governance. While you certainly could curtail some of the abuses by going to a user-fee system, remember the compromises that go with that. Other countries have shown that competent public governance is possible.

However, the more I think about it, the more I’m in favor of getting the federal government out of the capital projects side of things. Our mainline freeway and rail networks are complete, and the federal government seems to make a lot of poor investment choices now that most of the good capital projects are complete. There’s definitely an equity case for some federal involvement in helping out poor states and cities with operating costs and vehicle procurement, and the federal government should help states and cities out by using its low interest rate to borrow, but should the feds be involving themselves in things like Portland’s streetcar extension or the 69 freeway? Probably not.

Are Roads a Public Good?

You could make an internally consistent argument that drivers should pay the full cost of roads if you think that roads are not public goods.

I’m not buying that argument for arterials and neighborhood streets, since having two competing road networks in a city would be a huge waste of land, like having competing gas or electric companies. If arterials and streets were privately owned, they’d have to be regulated like a utility, and you’re right back to the issue of competent governance.

The argument is believable in the case of limited access tollways, where it’s easy to control access at onramps and offramps, and easy to manage demand through variable tolls. If public arterials are available, no one needs to use the freeway. However, I think there are practical limits to that model as well, which I’ll address in a separate post.

How Does Sustainability Impact Mortgage Defaults?

A new University of Arizona report on the effect of sustainability on mortgage default rates has been making the rounds of planning websites and blogs. The study was sponsored by Fannie Mae, so naturally it is concerned with improving the ability to predict mortgage defaults. The report concludes that sustainability features “may be used to improve the prediction of mortgage default”.

To its great credit, the report does not conclude that the presence or absence of the sustainability features have causative power over default rates. This tends to get lost in translation: while the report states that “properties with certain sustainability factors are a better risk than previously thought”, The Atlantic Cities makes the more nebulous statement that “every additional minute of commute time raises the risk of default 3.7 percent”, and on Twitter this is reduced to “proximity to a freeway/commute time increases default risk”.

Not quite. There is a subtle but important difference between saying sustainability factors “may be used to improve the prediction of mortgage default” and saying that lack of sustainability factors “increases default risk”. The fact that a variable has predictive power does not mean it has causative power, because it may just be a proxy for something else. The former suggests that someone underwriting mortgages should include sustainability factors in their model – something Fannie Mae might want to do. The latter suggests that we could improve society by building more mixed-use development  near transit and tearing down freeways – something The Atlantic Cities might want to do.

The question, therefore, is if sustainability factors have causative power over mortgage default rates. Since many people would consider the sustainability factors to be desirable neighborhood attributes, the obvious possibility here is that people with more choice self-select for those neighborhoods, while people with less choice are forced to accept longer commutes, less mixed-use, and more proximity to freeways. This is something that could be revealed through borrower characteristics. What does the report have to say about those?

“Relevant variables include borrower character, experience, financial strength, and credit history. Unfortunately, data on these issues were not available for this study.

It is unlikely that the omission of borrower characteristics as controls weakened the results. In linear regression, omitted orthogonal variables that are determinants of the dependent variable do not bias the parameter estimates… there is no reason to think that borrower characteristics would be correlated with the sustainability variables…”

In plain English, this means that the report did not include borrower characteristics, but that the authors do not see this an issue, because they do not expect any relationship between borrower characteristics and sustainability factors.

That assumption seems questionable to me. It is entirely plausible that disadvantaged borrowers – people in worse financial condition or with bad credit, not to mention groups that face housing discrimination – would end up in less desirable neighborhoods. In fact, there is a long history of this happening. None of this is intended to question the desirability of and the multitude of benefits that come from mixed uses and transit accessibility. But I don’t see how you can conclude from this report that building more TOD would have a meaningful impact on mortgage defaults.