Austin Contrarian

Trading Places

From Wells Dunbar, writing at KUT News:

It’s no secret Interstate 35 congestion takes a toll on Austinites.

Out of a list of Texas’ 100 most congested roadways, the portion of I-35 running through central Austin is the fourth most congested in the state. Meanwhile, State Highway 130, out east of I-35, is open for business.

Officials have tried all manner of incentives to divert traffic from I-35 to SH 130, including the lure of an 85 mph speed limit.

But SH 130 is tolled and I-35 isn’t. That has some folks asking if we’re tolling the wrong road.

Like me.  (I'm quoted.)

Some Legislators are inching closer to this. Representative Paul Workman recently proposed making SH 130 toll free, using the state's rainy day fund to make up the lost toll revenue.

I say "inching," but that's halfway there, right?

Eliminating tolls on lightly-traveled SH 130 in order to reduce congestion on badly-congested I-35 would make everyone who drives either route better off (or at least no worse off), but of course at a pretty steep loss of revenue to the state government. Eliminating tolls on SH 130 and imposing variable tolls on I-35 would maximize aggregate welfare (mainly by converting wasted time into useful money), but some drivers would be better off and some worse off.  Not surprisingly, it's the "make SH 130 free" half of the plan that gets proposed first.

Routefacts.org

Austinites for Urban Rail Action (AURA) has a new page up: Routefacts.org, a "one-stop repository of the key facts in the Austin urban rail sequencing debate." The idea is that there should be something approximating a public debate on the initial alignment of Austin's urban rail line, which in turn requires careful sifting of facts and data.

Of course, AURA's main point -- a point Routefacts.org makes quite effectively -- is that that the facts are mostly unknown and the data are mostly undisclosed. For example, we have ridership estimates for the proposed Mueller line, but not Guadalupe/Lamar. Advocates who want a robust public debate on alignment are left to fend for themselves using bus ridership numbers or estimates for the 2000 light rail route that are now over 13 years old.

 

Austinites for Urban Rail Action

Jace Deloney and Julio Gonzalez Altamirano have formed Austinites for Urban Rail Action (AURA) to advocate for a "successful, open and transparent urban rail process." Follow AURA on Facebook. Or twitter. (Or both.)

The impetus for AURA is the route selection for Austin's initial urban rail line. Should the initial line go up Guadalupe and Lamar, the route proposed in 2000, or instead run out to Mueller?

I'm no transit expert, but I vote Guadalupe/Lamar. Guadalupe/Lamar has more existing, proven demand (i.e., much higher bus ridership). The Guadalupe/Lamar corridor is basically at capacity right now, so light rail in reserved guideway could increase the corridor's throughput. And Guadalupe/Lamar has better geometry -- a long, straight line criss-crossed by most of the important east-west corridors north of UT. A Guadalupe/Lamar line could serve as the backbone for north Austin's mass transit system. A line to the Mueller Town Center couldn't. 

I gather that AURA has not been created to advocate a specific route, though, but to advocate a public, data-driven route-selection process. We haven't had a public, data-driven discussion of the alternatives yet, and it's not clear that we're going to get one. AURA aims to fix that.

Along these lines, here is Julio's excellent post on the need for more data.

The coming robocar apocalypse

How will driverless cars affect the built environment? A lot of the discussion lately has focused on whether driverless cars will affect the demand for rail. Michael Lind thinks they will make rail obsolete. Alex Block, sensibly noting that a change in technology won't change the relevant geometries, disagrees.

Matt Yglesias (responding to a Randal O'Toole post that has been taken down) foresees a parking bubble:

[R]ight now every metropolitan area in the United States contains many, many more parking spaces than automobiles. When you're at work, the space allocated for your vehicle at home sits there empty. When you're at home, the space allocated for your vehicle at the office sits empty. Malls build parking to accommodate demand during peak hours, and the spaces mostly sit empty off-peak. But if the cars could drive around without a human pilot, there'd be no need for such lavish supplies of vehicle storage. In principle, a metro area could get by with fewer than one parking space per car since even at minimum-demand times a nonzero quantity of vehicles would be in use. That's probably extreme, but right now depending on how you count we have somewhere between three and eight parking spaces per car. If the cars don't need to sit idly waiting for you until you want to leave (imagine a world of cheap, ubiquitous taxis) that number is going to become totally ridiculous. After exploding for about 60 years, the torrent of parking construction is going to halt very suddenly and then start shifting into reverse. That should even make some rail lines more useful. Commuter rail stations, for example, will no longer need to choose between park-and-ride and transit-oriented development models. Every station will be a little TOD neighborhood, and people from further away will get dropped off and picked up at the station without needing to worry about storing a car there.

I agree with Matt that if robocars become ubiquitous, then we will see a reduction in demand for parking. But we will not see a reduction in demand for vehicle storage.  Vehicles will continue to exist, to occupy space, whether someone is behind the wheel or not. Driverless cars will simply encourage a shift from static vehicle storage (a/k/a "parking lots") to dynamic vehicle storage (a/k/a "streets").

Donald Shoup has harped (pdf) on the phenomenon of "cruising for parking." It is one of the hidden costs of underpriced street parking. Underpricing street parking causes every spot to be filled, which, in turn, causes drivers to circle the block, searching for an open spot or waiting for one to open up. Shoup surveyed a number of studies which found that, on average, 30 percent of the traffic flow consisted of drivers cruising for parking.  Underpriced parking, counterintitively, means congested streets.

There is a built-in check on cruising for parking, though: every driver eventually wants to stop and get out of the car. Drivers aren't cruising for fun; they are cruising in the sunny hope of finding a place to store their vehicles and get on with their business. If a particular part of town gets a bad parking rep, then a lot of drivers won't even bother going there. The fact that, today, drivers must tether themselves to their vehicles while in motion limits the amount of congestion from cruising.

With robocars, though, we'll witness a new phenomenon: cruising in lieu of parking. Rather than cruise Congress Avenue for fifteen minutes looking for a spot so I can get a taco at Chipotle's, I'll simply have my car drop me off at the door and command it to spend the next 30 minutes getting in everyone else's way while I eat. It's a pretty simple cost-benefit calculation: add the cost of parking to the cost of the anticipated time it will take to find a parking spot; if that cost is higher than the cost of gasoline, then just order the car to cruise.

It won't take a lot of people engaging in this practice to congest the streets. Just imagine, say, an extra 200 cars continuously cruising up and down Congress during the middle of the day. What's worst is that the people who will benefit from the practice will incur none of the costs. On the contrary, the more gridlocked the traffic, the less gasoline their cruising cars will burn.

In fact, this perfectly rational practice will probably be so harmful, so patently selfish, so despised that it will be necessary to outlaw it. Which means everyone will still have to find a spot for his car, driverless or not. Which means that, despite the title of this post, we might not see a robocar apocalypse after all, or a parking bubble, either (other than the existing bubble that local governments have created with underpriced street parking and mandatory parking minimums.)

Run urban rail up Guadalupe/Lamar

Jeff Wood at The Overhead Wire has been running a series on the history and politics of Austin's urban rail endeavors. Today he's posted an immaculate argument for running the line up Guadalupe and Lamar rather than out to Mueller.  The aerial photo below, which contrasts the proposed line (orange) and a Lavaca/Guadalupe/Lamar alignment using current bus ridership, makes the case succinctly:

Guadalupe and Lamar are at capacity and have been at capacity for decades.  Planners are shying away from them, evidently because they are too congested. But "being too congested" is the reason to make this massive investment in rail in the first place. Reserved guideway transit is the only way to significantly increase the capacity of these prime gateways to UT, the Capitol Complex and downtown.   Spending a couple hundred millioin on the four-mile segment from UT to Mueller is largely a waste at this point -- that's not a congested route, and it doesn't have much existing transit ridership.

A path to 20%

Cap Metro is touting its January ridership for the Red Line: it gave 32,000 rides last month, a 57% increase over January 2011. Cap Metro confesses, though, that it is recovering only 3.1% of its operating costs from fares. That’s dismal. As is its $33.98 subsidy per rider. But Linda Watson assures us, “We're running the agency like a business, we're making some good, tough decisions to make sure we're strong--not just today, but well into the future as well."

A 3.1% farebox recovery ratio and $33.98 per rider subsidy are simply unacceptable. What we need from Cap Metro is, as Watson promises, business-like behavior. This means a business plan for raising the FRR and reducing the per-rider subsidy for the Red Line. What I’d like to see, in particular, is a configuration of schedule, ridership and fares that gets the FRR up to 20%. Cap Metro knows what it costs to operate the trains. It should be able to write down some schedule, ridership level and fare structure that produces the necessary revenue. Or it should tell us there isn’t one; that’s something the public needs to know if true.

I normally would assume that there is some path to 20%. Trains are a high-fixed-cost but low-marginal-cost business. The marginal cost of carrying another passenger on a train that will be making a trip anyway is approximately zero (until you reach capacity). The marginal cost of adding another round trip should also be low, assuming the train would otherwise be sitting idle: I imagine the per-hour cost of an operator’s time is pretty low, and a train surely doesn’t use that much fuel for a 64-mile roundtrip. So the path to higher FRR should be a straightforward one of adding more fare-paying riders.

I’m not sure this is right, though. I say this based on the cost estimate Cap Metro quoted the city for adding Friday evening and Saturday service. It worked out to around $40,000 per weekend. If you assume an extra 40 runs per weekend, that works out to $1,000 per one-way, hour-long trip. Cap Metro reported an operating cost of $1,300 per hour in its last quarterly report (which you can find here), so this seems to be the right ballpark. (I don’t know why it costs Cap Metro so much. Cap Metro actually contracts with Herzog Transit Services to operate the train, so the costs, whatever they are, are built into that contract.)

We need to know if there is a path to a more reasonable FRR and per-rider subsidy, though. We need to know this because we ultimately have to decide on the optimal level of service for the Red Line. If there is a schedule, ridership and fare structure that gets the FRR up to a reasonable level (which presumably would entail a substantial boost to ridership) then that should be Cap Metro’s goal. But it might be that, given Cap Metro’s current cost structure, there is no ridership level or fare structure that will cover an acceptable portion of the operating cost. Perhaps the very best case scenario is that Cap Metro can bump the FRR to 10% only by tripling its annual operating loss. If that’s the case, then we don’t want Cap Metro to maximize its FRR, but rather to minimize its operating loss, conserving money for buses. That would mean sticking to a light schedule.

Either way, mere ridership numbers don’t tell us what we need. We need a plan.

I forgot about Cap Metro's interests

Last week, I concluded that the plan to add two tolled lanes to MoPac would make everyone better off than either maintaining the status quo or adding two free lanes.  But it occurred to me that I haven't considered everyone's interests.   Cap Metro's, specifically.  Isn't it likely that Cap Metro will be worse off?

As I noted last time, congestion-pricing will be a boon to bus riders.  They will get much shorter, more predictable rides.  But these buses run from the stations (Leander and Lakeline) that provide the bulk of Cap Metro's morning boardings.  Cap Metro is to some extent competing against its own buses for its rail riders.   (Cap Metro's surge in rail ridership last year occurred after it scaled back its express bus service from Leander and Lakeline.  The reduction in bus service might not have been the sole cause of the bump in ridership, but it surely played a part.)  

The shorter, more reliable bus trips will inevitably attract some current rail riders.  What's more, Cap Metro will be under a lot of pressure, I predict, to expand express bus service when the tolled lanes open.  The benefits for express bus riders are one of the big selling points of congestion pricing.  Free-flowing lanes will make it possible to offer reliable express bus service during the very the peak of the peak of the morning commute.  Cap Metro will be in the difficult position of either providing great, new express bus service at the cost of its rail ridership or refusing to provide great, new express bus service to preserve its rail ridership.

Perhaps it's not useful to frame this in terms of Cap Metro's interests.  Cap Metro's mission is to provide quality, cost-effective public transit.  If changes occur to the transportation network that allow Cap Metro to provide better public transportation, then we should count that as a win, even if it puts Cap Metro in a difficult position institutionally. 

So perhaps this is a better way to frame the point:  adding congestion-priced lanes to MoPac will put more pressure on the Red Line.  Cap Metro is running the Red Line today at an unsustainable subsidy of $33.98 per rider.  I don't know how Cap Metro plans to reduce that subsidy to a reasonable level.  I'm pretty sure, though, that improved bus service on MoPac will make the path more difficult. 

The Freeway Congestion Paradox

I've gotten some pushback on my claim that adding two congestion-priced lanes to MoPac will increase capacity more than adding two free lanes.

This is the Freeway Congestion Paradox:

Chen and Varaiya, in their article entitled "The Freeway Congestion Paradox," have demonstrated that, once freeway vehicle density (measured in vehicles per mile) exceeds a certain critical number, both vehicle speed and vehicle flow (measured in vehicles per hour) drop precipitously.  They have demonstrated the phonemenon with actual data from a section of westbound I-10 in Los Angeles.  Until 5:10 am, a flow of 2,100 vehicles per lane per hour is maintained, at a speed of 58 mph.  As density increases after 5:10 am, speed steadily drops, until at 7:00 am speed is a stop-and-go 15 mph, and flow decreases to 1,300 vehicles per lane per hour.

The Federal Highway Administration commissioned a study in 2008 that looked at data from Washington, DC freeways.  From the summary:

Highway studies have determined that once traffic volumes exceed the capacity of the roadway, the system can rapidly "break down" to the point where all traffic slows markedly, and the capacity and throughput of the roadway drops precipitously.

They produced neat graphs depicting the drop in capacity caused by stop-and-go traffic, like this one:

This chart pairs vehicle speed and total volume carried (see pp. 15-16 of the report for more discussion).  The dots are connected in temporal sequence, so it's possible to follow the evolution of the traffic jam.  The graph has a a random-walk component, but basically it confirms two things.  First, low throughput volumes are associated with low speed (i.e., stop-and-go traffic).  Second,  after the critical threshhold is passsed, it takes traffic a long time to recover. 

Here is the same chart, except I have drawn over some of the line segments in red.  This shows that once the tipping point is reached, traffic speeds and volumes drop and take a long time to recover.   (Just trace the red line segments starting at "A").

Congestion pricing would have prevented congestion from reaching the tipping point, effectively increasing the highway's capacity.

Rush-hour traffic on MoPac today is the stop-and-go 15 mph type.   Adding a new, free lane would probably shorten the period when the highway is clogged, but it would still be clogged for long stretches of time.  During these periods, a new, free lane would carry less traffic than a tolled lane.

When congestion is not bad, the toll should be low.  That's one of the attractive features of variable congestion pricing -- the price, in fact, should be zero when the marginal cost of congestion is zero.  (According to the Statesman, CTRMA might charge a nominal amount like a quarter to avoid confusing people.  Which confuses me.)  Put differently, the new lane should "act" like a free lane when traffic is not bad.

Congestion, envy and equity

The Central Texas Regional Mobility Authority plans to add two lanes to MoPac between Lady Bird Lake and Parmer Lane.  The cool part of the plan is that it intends to use dynamic congestion pricing -- i.e., the toll will fluctuate as necessary to keep traffic in the the tolled lanes flowing at 50 mph.   The six existing lanes will remain free.

This is how we ought to add new capacity.  It will make everyone better off.   The people who choose to pay the toll will be better off because they value the time savings more than the cost of the toll.  Bus commuters will be better off -- they might be the biggest beneficiaries, in fact -- because they will get a suddenly much shorter commute for (I presume) the same bus fare.  Drivers who continue to use the free lanes will endure slightly less congestion, even if it's just a narrower period of peak congestion.  Finally, taxpayers, if not better off, will be no worse off because they won't have to pay for the extra capacity.  The capacity will be paid by those who value and use it.

Everyone will not only better off compared to the status quo, but . . . and this is the key point . . . they'll be better off than if the two new lanes were free.   Adding two free lanes would reduce congestion, too, but they would reduce congestion less than two new tolled lanes.  Tolling congested lanes increases their capacity.  A congestion-priced lane can handle 1,800-2,000 cars per hour; an unpriced lane during peak congestion will handle less than half of that.  Somewhat counterintuitively, perhaps, charging for a congested lane is a sure way to get more cars through it.

Obviously, drivers happy to pay the tolls to skip congestion would be worse off if the new lanes were free rather than tolled.  Bus riders who would lose out on a faster commute would be worse off if the new lanes were free.  Taxpayers stuck paying for capacity that they might not value would be worse off if the new lanes were free.   But, because tolling lanes increases their capacity, even the drivers who would not use the toll roads anyway would be worse off if the new lanes were free.  They would have to contend with more congestion.   

Congestion pricing is a no-brainer.  So I was puzzled when Ben Wear raised the equity angle in a companion column.   After pointing out that all drivers would benefit from the congestion reduction from the new lanes, he concluded:

Undoubtedly however, those with money will be able to use the variable-price lanes more often than those on a budget.

It may become something of a spectator sport for those caught in MoPac congestion to count the number of Lexuses, Mercedeses and Escalades that fly by in the express lane.

It ain't necessarily fair. But in the words of John F. Kennedy (who certainly could have afforded to use this lane), life is unfair. And it will help pass the time.

It's strange to complain about fairness when everyone will be better off with new tolled lanes than new free lanes.  

But it's especially strange to complain about fairness because, once the lanes open, the only people on MoPac who will be causing a net social loss will be the drivers in the free lanes.

 This is basic congestion theory but most people still don't get it.  When you enter a congested highway, your mere presence inflicts a cost on everyone else.  Every car behind you is delayed just a little bit, maybe only a few seconds each, but the delay affects a bunch of cars.  Getting in everyone else's way imposes a social cost.

There's not a lot a driver can do about that today.  We can't even say for sure what cost each driver is imposing on everyone else.  But once the toll lanes are built, we'll know.  We'll know because the congestion cost each driver imposes on his fellow drivers will be approximately equal to the price of driving in the congestion-free lane.  That's one of the quirks of the economics of congestion -- it turns out that the cost of congestion a driver imposes is equal to the congestion-clearing price.

Once the toll lanes are built, drivers will have a choice.  The people who enter the tolled lane will be precisely the drivers willing to pay the cost of their congestion.  The toll lane will allow them to pay that cost and get out of the way of everyone else.

The people who don't pay the cost are people for whom, say, 15 minutes in time savings isn't worth the $3 toll or whatever it turns out to be.  That's most people, presumably.  But saving 15 minutes is worth something to them.  If you value the time savings at $1, you would incur a $2 loss by taking the tolled lane.  But you will impose a $3 loss on your fellow drivers if you stay in the free lanes, because we know the cost of your congestion is equal to the toll.  Staying in the free lanes produces a net social loss of $1.

Congestion is socially-sanctioned selfishness.  It's ok only because everyone does it.  But if we'd get used to thinking in terms of costs we're inflicting on everyone else, there'd be less moaning about those paying for the cost they're imposing while we mooch our way down the free lanes.  

Sometimes parking regulations don't matter

I think minimum parking requirements are bad policy.  If I had my way, the City would abolish its minimum parking requirements tomorrow. 

I'm a Shoupie.

But it's worth reminding oneself every now and then that the regulations on the books -- even the bad ones -- don't always matter. Minimum parking requirements don't matter when the market demands (or, at least, the developer perceives the market to demand) more than the minimum.

Greystar's proposed mixed-use project at 1200 S. Lamar (the South Alamo drafthouse site) is a case in point. According to the parking calculations on the site plan, the project will provide almost 60% more parking than code requires.

I'm sure this demand for parking is being driven by the Alamo Drafthouse. Anyone who's been to the South Lamar Alamo -- or tried to get take out from Suzi's China Kitchen -- on a Saturday night knows that the Alamo gobbles up parking.

There is good evidence that parking regulations do matter in general. They can matter a great deal in specific instances. How much extra space do we reserve for empty asphalt thanks to city regulation? We don't know and we can't know unless the city gets rid of its minimums. But it's important not to be seduced into believing that every massive parking garage or expansive surface lot is a manifestation of bad city regulation.