Coyote Blog

The previous annual installment in this series was here.

Well, the Phoenix light rail system has hit new levels of fail that even I could not have projected.  Here are the annual numbers updated for the most recent report:

Yes, I know that supporters will argue that the agency should get a "pass" due to COVID.  I am willing to do so for the bus system, but not for light rail.  Let me explain.

I have always been, perhaps unlikely for a libertarian, comfortable with some base level of public transportation.  Lack of mobility can be a huge barrier to upward mobility and I can support public transit systems in major cities as a sensible anti-poverty program.

What I oppose is the replacement of relatively inexpensive and tremendously flexible bus transit systems with light rail.  Light rail requires orders of magnitude more capital and operating cost per passenger (except perhaps in the absolute largest and densest cities, which does not include Phoenix).  Light rail is limited by time and money and necessity to only a few routes.  In almost no city outside the centers of the few largest cities (think Manhattan) has it ever been possible to develop a complete enough web for citizens to entirely give up a car.  And light rail is absolutely inflexible -- once capacity is added on one route (for literally billions of dollars) it can't be moved where busses can be shifted easily from route to route.  For example, on a days' notice we had bus routes going to the major local vaccination centers -- no way to do that with rail.

It has always seemed to me that light rail is a middle class boondoggle.  White professionals kind of like the trains whereas they won't ride busses.  The few light rail routes tend to follow middle class / professional commuting routes.  Light rail gets professionals from home to office (after which they use their car for other sorts of trips) but does little to provide mobility for the poor to the doctor, to the store, to support services, etc.  Worse, because light rail costs 10x or more per passenger than busses, the advent of light rail generally starves a transit agency of funds and causes them to cut back on busses to save money (because once the rail is laid, you can't really cut back on it).  Look at the chart above, despite years of light rail extensions, total transit ridership in Phoenix stalled after growing for over a decade.  Wealthy university students at ASU got a nice train to the downtown clubs, but the poor who depend on busses for basic mobility lost a lot of service and routes.

I will confess that I didn't anticipate a pandemic in my prior opposition to light rail, but it has been more proof of the light rail fail.  Busses that are not needed are parked, sidelining a modest capital investment.  But light rail is $2+ billion in capital investment lying dormant.  With the middle class eschewing transit during COVID, transit is returned to its core function of providing essential mobility to the poor.  But it can't do this as well as it used to because of the multi-billion dollar light rail capital albatross.

Well, it is that time of the year when Valley Metro, the Phoenix area transit authority, publishes its fiscal year (July-June) ridership numbers, and boy are they awful this year.

In past installments, I have pointed out how our shift to funding a ridiculous light rail system in one of the least dense cities in the world has killed transit ridership.  Since rail costs more than 10x per passenger mile to move riders vs busses, the shift to light rail tends to starve bus funding in every city it has been tried, leading to net losses (even pre-Covid) in total ridership numbers -- yes even in Portlandia.  This is all discussed more in our 2019 transit report.  Here is a brief bit:

The problem with light rail (and the reason it is popular with government officials) is that it is an upper middle class boondoggle.  There can be no higher use of transit than to provide mobility to poorer people who can't afford reliable automobiles.  Buses fulfill this goal better than any mode of transit.  They are flexible and can reach into many corners of the city.  The problem with buses, from the perspective of government officials, is that upper middle class people don't like to ride on them.  They like trains.  So the government builds hugely expensive trains for these influential, wealthier voters. Since the trains are so expensive, the government can only build a few routes, so those routes end up being down upper middle class commuting corridors.  As the costs mount for the trains, the bus routes that serve the poor and their dispersed commuting destinations are steadily cut.

Without further delay, here are the updated numbers including 2020. (Tellingly, Valley Metro has for some reason failed to update its chart on its ridership web page to include 2020 numbers.   I wonder why?)

As you can see, even before COVID we were trending towards a total system ridership decline vs. the days before the commitment to light rail.  And we certainly had killed the growth in system ridership we had seen prior to light rail.  But now, with 3-4 months of COVID in the numbers, ridership has fallen off the map.  Boy, I sure bet we would be better off right now with a bus only system where capacity is easier to add, subtract, and relocate vs. light rail whose costs and routes are literally sunk in the ground.  I have no opinion on whether transit ridership will return to pre-COVID levels but I am sure that such massive disruptions in the system are far easier to economically manage in a bus network than in a rail network.

But you will be happy to know that Valley Metro is still spending money to put rails in the ground -- in fact, it has accelerated its spending post-COVID if you can believe it:

Phoenix is ramping up its light rail system despite record-breaking declines in ridership spurred by COVID-19.

Light rail ridership is down a little over 50% from where it was pre-COVID-19, according to Valley Metro CEO Scott Smith. Bus ridership is down about 40%.

Light rail construction hasn't slowed, however. In fact, Valley Metro has accelerated some projects, taking advantage of less vehicle traffic on the road while more people stay home.

Does this make a bit of sense?  Here is the defense:

But with ridership at all-time lows and no indication as to when ridership will return at pre-pandemic levels, is it smart to continue with pricey expansion of the light rail system?

"I think if we used that same logic, we'd say 'Let's go tear down all the movie theaters and all the stadiums.' Nobody expects this to be around forever. We're going to at some point in time, whether it be in the next month or the next year, return to some normalcy," Smith said.

Ooh, @Scott_SmithAZ has obviously had an expensive education that includes lessons in how to defend a bad position using logical fallacies.  Comparing expanding light rail to tearing down movie theaters is a false analogy.  No one is talking about tearing out light rail (well, maybe except for me).  The appropriate comparison is to ask if anyone is saying "let's go build a bunch of new movie theaters and sports stadiums" right now, and the answer is definitely no.

But don't despair, Phoenix taxpayers.  Because Valley Metro is clearly carefully considering its routes and is only building those to locations they are sure will be super popular both during and after COVID.  For example, one current expansion is to connect the existing network to... an aging shopping mall.

To the surprise of no one who reads this blog, Phoenix Light Rail continues to fail at every goal it set for itself, despite over $2 billion in cumulative expenditures and over 10 years of time to "catch on."  In fact, the only goal it has succeeded in is to allow our local leaders to virtue signal with their peers that they have a "modern" transit system, "modern" for some reason defined as increasing the use of a 150-year-old transit technology.

Here are the numbers, relatively fresh from Valley Metro's ridership reports.  Valley Metro has stopped (for reasons likely to become clear in a moment) publishing long-term ridership trend charts, so I have made my own from their data.

The original Phoenix light rail line cost about $1.4 billion.  The one line was expanded midway through the 2016 fiscal year by 31% in length and 36% in cumulative dollars spent (to very little obvious effect).  To date, well over $2 billion has been spent on the line.

What we can see here is exactly what Randall O'Toole of Cato has been saying for years -- that light rail projects tend to actually hurt total transit use as they scavenge resources from other modes, like buses.  This is because light rail costs so much more to move a passenger, both in terms of capital investment and operating cost, so $X shifted from buses to rail reduces total system capacity and ridership substantially.  We have seen this in Phoenix, as light rail costs have forced closing or reduced services in a number of bus routes, with obvious results in the ridership numbers.

There are a couple of ways to look at the ridership numbers and our $2+ billion light rail investment.  First, this investment has gained us an additional 1.25 million round trips on the system, at a capital cost of $1600 per added annual round trip ($2B divided by 1.25MM).   For someone who commutes 250 days a year, this means the Phoenix taxpayer has spent $400,000 per incremental daily commuter on the system.  I used to predict that we could more cheaply buy every rider a Prius with the money we spent on the system, but I obviously underestimated how bad it was.

Another way to look at this is that in fact we did not increase ridership at all, but decreased it.  From 1997 to 2008 the system increased ridership by an average of 5.6% a year.  Had the same growth rate obtained for the last 11 years since light rail was put in place (and certainly Phoenix has continued its fast growth in this period) the ridership of the system should have been north of 110 million trips a year.  By this metric, we spent $2 billion to shift to a costlier, less flexible transit mode that lost us 45 million annual trips.  Even if transit had just grown with population we should have seen 80 million or so trips in 2019.

The problem with light rail (and the reason it is popular with government officials) is that it is an upper middle class boondoggle.  There can be no higher use of transit than to provide mobility to poorer people who can't afford reliable automobiles.  Buses fulfill this goal better than any mode of transit.  They are flexible and can reach into many corners of the city.  The problem with buses, from the perspective of government officials, is that upper middle class people don't like to ride on them.  They like trains.  So the government builds hugely expensive trains for these influential, wealthier voters. Since the trains are so expensive, the government can only build a few routes, so those routes end up being down upper middle class commuting corridors.  As the costs mount for the trains, the bus routes that serve the poor and their dispersed commuting destinations are steadily cut.

Some of my other posts on Phoenix light rail are here.  As you can see from this 2008 article, I am not just a Monday morning QB but predicted exactly this failure in advance and explained why from data.

The only thing more annoying in an argument than trying to have a discussion with someone who does not think logically is to have an "ally" pop into the discussion on your side who does not think logically.  Via the AZ Republic

One of the original leaders of the anti-light rail movement in south Phoenix claims God "judged" Congressman Ed Pastor for "bringing death" to the community by supporting light rail and punished him with a fatal heart attack.

At a City Council meeting last week,Celia Contreras told council members she was "coming in the name of Lord Jesus Christ" with a message: Stop the light rail or the "punishment" will continue

Well, to be fair, the guy knows his business is going to take a big hit, without any hope of a takings claim on the government, and I suppose he is pretty stressed out.

By the way, the presence of this story in the Republic is a tell as to which side the paper favors.  I have been to public meetings on Phoenix light rail and I have personally seen a number of insane claims by light rail supporters (at lot of wrath of Gaia stuff, for example) that never gets featured in the paper.

The other day I got stuck in Orlando for a day and took the opportunity to sit it out at DisneyWorld.  Despite being essentially a walk-up guest, the hotel upgraded me to a beautiful room looking out at the Magic Kingdom and the lake in front of it.

The Magic Kingdom is an interesting public transit case.  It is located miles away, across a lake, from its parking lot and can be quite a distance from many of the Disney hotels.  So most everyone comes to the Magic Kingdom in some sort of mass transit.

The most eye catching is the DisneyWorld monorail.  This version, designed in the late 1980's but fairly similar in outline to the original early-70's version, is clearly beautiful.  When I talk about industrial design, I often use a scale where 1 is the Boston City Hall and 10 is the Disneyworld monorail.  Not only are the trains themselves beautiful, but the spindly monorail beams are clearly better looking than most any other overhead rail arrangement.

But what makes for a good theme park ride does not necessarily make for good public transportation.  Here is why:

1. Monorails like this give up all the efficiency of trains.  Rail is efficient because the rolling friction of a flanged metal wheel on a metal rail is so much lower than, say, a tire on concrete.  But these monorails and most others ride on tires -- there is rolling friction of the tires on the top of the concrete beam as well as with the tires that stabilize the train on the sides of the beam.  These monorails are basically long busses up on a rail.
2. Monorails have lower capacity per car than trains.  For stability reasons on the narrow beam, monorails must be shorter and narrower than most rail trains (even given the world's too-narrow rail gauge standard).  This means for roughly the same car length and weight, they carry fewer passengers.
3. Monorails have the same downsides of trains.  They have capacity restrictions -- specifically they need much longer headways between trains for safety than busses do on the road.  And they are inflexible -- once you spend billions of dollars to put them on one route they are going to stay on that route.

As I sat on the lake, I could see all three modes of transit that Disney uses at the Magic Kingdom.  Both monorails and high capacity double-deck ferries run from the parking lot to the park, and busses run from most of the hotels to a bus depot at the park.  I could be wrong because it was not scientific, but I believe that the ferry boats had a lot more capacity per hour than the monorails -- particularly since Disney had to lengthen the distance between trains several years ago after an accident (yes, they actually have a single loop of track with no switching a couple of miles long and with perfect visibility ahead and they actually had a collision).

More telling is the fact that Disney has essentially abandoned the monorail for all its future expansions.  Yes, they built one to Epcot, but they have reduced service to one every 15 minutes or so and it brings in an absolutely trivial portion of the guests.  With its growing network of parks and hotels and all the possible point to point destination pairs, it relies on busses now almost exclusively for internal transport.

Megan McArdle has a great WaPo article and tweet storm on high-speed rail in the US.  In it she focuses on issues of distance and infrastructure barriers we have.

One thing she left out is that the US rail system is optimized for freight, vs. European and Japanese systems that are optimized for passengers (it is hard to do both well with the same network).  The US situation is actually better, much better, for energy conservation.  I wrote in detail about this before:

First, consider the last time you were on a passenger train.  Add up the weight of all the folks in your car.  Do you think they weighed more or less than the car itself?  Unless you were packed into a subway train with Japanese sumo wrestlers, the answer is that the weight of the car dwarfs that of the passengers it is carrying.    The average Amtrak passenger car apparently weighs about 65 tons (my guess is a high speed rail car weighs more).  The capacity of a coach is 70-80 passengers, which at an average adult weight of 140 pounds yields a maximum passenger weight per car of 5.6 tons.  This means that just 8% of the fuel in a passenger train is being used to move people -- the rest goes into moving the train itself.

Now consider a freight train.  The typical car weight 25-30 tons empty and can carry between 70 and 120 tons of cargo.  This means that 70-80% of the fuel in a freight train is being used to move the cargo.

This is another case of short-sighted analysis that looks only at the seen rather than the unseen.  Coastal elites take trips to Europe and see the beautiful high-speed trains and in turn never spend a moment thinking about freight trains.  So they fixate on beautiful sexy passenger trains rather than thinking about the system holistically.  I titled a Forbes article I wrote on the effect as "Shifting Capital from the Productive to the Sexy."

PS-  I am a train buff and have a whole room of my house filled with a model railroad, so I don't knee-jerk hate on rail.  I have ridden European high-speed rail many times so I am familiar with the product.  The London-Paris segment is great, and I have ridden the French TGV from Paris to Marseilles and the Italian line from Milan to Florence.  What's not to love as a tourist -- we don't pay for them and they provide good service between the city centers of tourist destinations.   But if you look at those trains they really have a ton of expensive infrastructure carrying not very many people over relatively (for the US) short distances.

I write this because after I criticized infrastructure triumphalism in Joel Epstein's article at Huffpo, he wrote me a one line retort: "You should get out of the country more often."  LOL, if you had to enshrine a hall of fame of sneering coastal elite dismissive comments of critics, this would have to be on the list.  I tried to follow up with him and ask him if he would have the US adopt China's infrastructure construction practices if the cost was adopting China's environmental and accountability standards, but I did not get a response.

A few weeks ago I posted Valley Metro's own numbers that showed that the billions spent on light rail in the Phoenix area have done nothing but case the stagnation of transit ridership in Phoenix.  Light rail ridership actually fell substantially despite an expensive extension of the line.

Today, the Antiplanner has an update on light rail in Phoenix, and it is not pretty:

As of 2016, light rail carries less than 0.2 percent of all travel in the Phoenix urban area. The 2016 American Community Survey says that the same tiny percentage of commuters take light rail to work, which is unusual as transit’s commuter share is usually much higher than its total share. Phoenix light-rail ridership in the twelve months ending in June, 2018 was down 4.4 percent from the previous twelve months. Transit ridership for Phoenix as a whole is down 5.6 percent for the same time period.

Phoenix is one of many Sunbelt urban areas in which rail transit makes no sense at all. Aside from the Antiplanner’s argument that buses can move more people than light rail, rail systems only make sense where there is a high concentration of downtown jobs that a hub-and-spoke transit system can serve. According to Wendell Cox’s calculations, downtown Phoenix has only about 26,000 jobs, which is just 1.4 percent of jobs in the metropolitan area.

Phoenix is particularly unusual (though not unique) in that its suburbs are actually denser than the city itself. According to the 2010 census, the city of Phoenix has about 2,800 people per square mile, while its suburbs have nearly 3,500 people per square mile. With both jobs and population spread out, the region needs nimble, low-capacity transit if it needs any transit at all.

Arizona State University students make up a “substantial component” of light-rail riders. Until this year, students were able to buy transit passes for 200 for the nine-month school year, plus $75 for the other three months. The same passes would cost other members of the public $768 per year. Despite the steep discounts, student weekday ridership dropped by around 40 percent between 2011 and 2015.

The last paragraph reminds me that I forgot to discuss the issue with ridership and ASU students in my last post.  A huge portion of Phoenix light rail ridership comes from two sources:  subsidized ASU students and fans going to downtown sporting events.  This helps to explain why the commuter share of Phoenix ridership is so low -- essentially, many of the light rail riders are not commuters but in these two other groups.  Why we should be spending billions to subsidize bar crawls for already heavily-subsidized ASU students or to save sports fans money on downtown parking** is beyond me.

He has this good news:

The Phoenix city council is considering delaying or even killing some planned light-rail lines because it is concerned that city streets are falling apart and too much money is being spent instead on an insignificant form of travel.

** much of the downtown parking revenue for sporting events goes to the city and county, so cannibalization of this revenue is yet another hidden cost of light rail.

Well, another year's ridership numbers are out for Valley Metro and Phoenix light rail and they are just as grim as they have been every year since Phoenix spent the first $1.4 billion on the first leg of the rail system (source)

Now, this picture is bad enough, until you realize that Valley Metro completed a huge extension of the rail line in 2016.  In 2016 the line length was increased by 31% and the cumulative capital investment increased by 36%.  With, as you can see, essentially zero effect on rail ridership in red.  The only small highlight was that after falling for years, bus ridership actually perked up a few percent.  As you may remember from earlier posts, bus ridership could be expected to fall due to cannibalization from light rail, but in fact it tends to fall even faster than rail ridership rises, causing total ridership to fall.  The reason is that light rail costs at least an order of magnitude more (including amortized capex) per passenger mile than busses, and so light rail tends to starve the bus system of funds.  Every light rail system implementation has been met with the need to slash bus service to pay for the huge light rail costs.  So despite enormous operating subsidies and more than $2 billion in cumulative capex, rail ridership has been flat and total transit ridership has fallen.

But in fact the picture is worse than this when you look over a longer timeframe, which is why Valley Metro has probably changed their practice from graphing nearly 20 years of history to graphing just 6.  Here is an older chart of theirs I posted years ago: (the top year in this chart is the bottom year in the chart above)

I will get back to the annotation in a moment.  But notice that despite all the cost and disruption and higher taxes from the light rail system, total ridership this year of 66.8 million is less than any year since light rail was opened and baredly 8% higher than it was before light rail opened 10 years ago in 2008.  Just organic city growth and recovery of the economy since 2008 should have driven faster growth than this.  In fact, in the 10 years before light rail was opened, Phoenix transit ridership grew 70%.  If that organic growth rate in bus service had been allowed to continue without the backbreaking costs and limited capacity of light rail being added to the mix, we should have expected 105 million riders this past year, not 66.8 million.

The Union Pacific Railroad is attempting to restore and run a 4-8-8-4 Big Boy locomotive next year.  Seeing live steam locomotives run is a rare but fabulous spectacle, but most of what is still running is WAY smaller than this locomotive they are restoring.  I am not sure most people have a sense of scale for these beasts.  I chose the picture below because it has people in them for scale.

This article on an expansion of the already grossly under-performing Seattle trolley system has everything:  Over-optimistic ridership and revenue projections and no better service than busses while costing substantially more.  It's hard to figure out where to even excerpt this article, which is the kind of skeptical media coverage of light rail and trolleys that one almost never sees, particularly in a Progressive city like Seattle.

Your fares cover about 40 percent of operating costs for Sound Transit’s Link light rail. Fares cover about 31 percent of the cost of King County Metro buses. Seattle’s two streetcar lines cover 23 percent of their costs with fares.

But once a streetcar is built along First Avenue in downtown Seattle, the city Department of Transportation (SDOT) expects fares to cover a whopping 56 percent of operating costs for the three lines.

That would be among the highest rates of any transit agency in the country.

And it’s one of a number of optimistic financial projections contained in Seattle’s plans to expand a streetcar system that is performing far below expectations.

“The financial assumptions are simply unrealistic based on our history with the streetcar,” Seattle City Councilmember Lisa Herbold said. “I don’t want a situation where we don’t meet those projections and the result is we end up seeing bus-service hours cut to pay for any shortfall.”

Which is exactly what happens with every trolley and light rail project, as I reported on in Phoenix recently.

The two streetcar lines — South Lake Union and First Hill — have low ridership that the city doesn’t expect to improve until the First Avenue line is built. But once it is built? The city predicts an exponential surge.

So far this year, the city’s two streetcar lines have averaged about 5,200 riders per weekday, if you filter out 2½ weeks in the spring that the First Hill line was shut because of an uncontrolled skid.

Those numbers, for the two lines combined, equate to about the 25th-most-popular bus line in King County.

SDOT predicts that in its first year of service, the expanded streetcar system will more than quadruple its ridership — to nearly 22,000 weekday riders....

SDOT officials say their ridership model is the gold standard — used by the Federal Transit Administration to validate other agencies’ models.

I loved that last line - "used by the FTA"... not mentioning this model has been overestimating ridership at every other project in the country. I would say it is the gold standard for producing figures to try to sell a bad project to taxpayers.

I wish every article on light rail disclosed this critical piece of information:

Westlake Avenue is a microcosm of the arguments made by streetcar skeptics nationwide, who question their utility if they are not separate from other traffic.

“If you build a streetcar instead of a good bus line, that money you spend above the cost of the bus line is not helping anyone get anywhere faster,” Jarrett Walker, a Portland-based transit consultant, wrote in 2009.

Bus lines are almost always cheaper to build than streetcar lines and they’re also less expensive to operate. It costs King County Metro about $163 an hour to operate a bus, according to federal data, while a South Lake Union streetcar costs $242 an hour, although a streetcar can carry 30 to 40 more people than a double bus.

I will end with the line my reader pulled out from the article

Why, many transit users realized, should they ride the streetcar, which runs barely 1 mile, when buses come four times as often, cover the same stretch as quickly and also extend their routes to West Seattle and Ballard?

Well, the numbers are in for the 2017 fiscal year (which ends June 30) and after another huge investment in light rail, Phoenix has lost more transit ridership.   From the Valley Metro web site:

First, credit where it is due.  After years of bizarre chartsmanship where bars on their graphs bore only a passing relationship to the numbers being graphed, Valley Metro seems to have adopted a new (or their first) graphing program.

As you can see, while light rail trips were up by about a million, bus trips were again down by over 2 million, for a net loss in transit ridership of over a million, the fourth year in a row this has been the case.  I had expected rail ridership to rise, since in 2016 the rail system was expanded by 31% in length and 36% in cumulative investment.  This extension resulting in a 15.6% increase in rail ridership between 2015 and 2017.  Early on, I got in a debate with supporters of the line arguing that since they had cherry-picked the densest corridor in town to start, incremental extensions would actually reduce ridership per mile because they would be serving less promising routes.  Supporters argued that I was ignoring network effects and that ridership would rise faster than line length.  I guess we are sorting out that argument now.

In the ten years leading up to the opening of the light rail line, transit ridership grew by an average of 6.7% a year in Phoenix.  In the 8 years since the rail line's opening, total transit ridership has fallen 1% per year.  This is a well known effect (at least well known to all but rail die-hards) that Randal O'Toole, among others, has been pointing out for years.  Since light rail is an order of magnitude more expensive to operate per passenger-mile, and since transit budgets are never infinite, growing light rail tends to strangle bus traffic, because bus routes and service have to be cut to feed money into the light rail money pit.  Since every dollar spent on rail moves fewer passengers than a dollar spent on buses, transferring money from buses to trail reduces total ridership.  It is worth noting that had the line not been built and bus transit had been allowed to grow as it had before the line, there might have been over 40 million more trips last year assuming pre-2009 growth rates.

As someone whose company engages is what is usually called "public-private partnerships" or PPPs, one would expect me to be an enthusiastic supporter of all such efforts.  (As an aside, my company privatizes the operation, but not the ownership, of public parks and we are paid entirely by user fees and get not one single dollar of tax money.)

But I totally agree with Randal O'Toole's frustration here, talking about light rail in Denver:

Now RTD has been forced to admit that two other lines being built by the same company won’t open on time. RTD claims that it saved money by entering into a public-private partnership for the line in what is known as a “design-build-operate” contract. In fact, it saved no money at all, but was merely getting around a bond limit the voters had imposed on the agency. If the private contractor borrows a billion dollars or so and RTD agrees to pay the contractor enough to repay the loan, the debt doesn’t appear on RTD’s books. Taxpayers will still end up paying interest in the loans, which actually makes it more expensive than if RTD had stayed within its debt limit.

Public-private partnerships work great if the private partner is funded out of the user fees collected for the project, such as a toll road or water system. The Antiplanner resents the way the transit industry has coopted the term, public-private partnership, because their kind of partnership works differently. Instead of being dependent on fares, the private partner gets a fat check from the agency each month–up to $3.5 million in this case–whether anyone rides the train or not. This means the private partner has little incentive to make sure the system is working. RTD has withheld a portion of the monthly payments until the problems are solved, but eventually the contractor will get all of the money.

The solution isn’t for the agencies to build the lines themselves. The solution is to completely avoid megaprojects that aren’t funded out of user fees. Without the discipline of user fees, everything that’s happening with the A line should have been expected.

Valley Metro, the agency that operates light rail and most bus service in the Phoenix area, has published its 2016 annual ridership numbers, and they are awful (the agency is on a july-june fiscal year).  Over the past year, they have opened two extensions of the current light rail line, spending $o.5 Billion to extend the line 6.2 miles.  So what did we get for this?  Falling transit use.

To begin with, we must again look past Valley Metro's downright bizarre chartsmanship, where differences in bar length bear absolutely no relationship to the values graphed (just try to figure out the bar lengths for the last three years of light rail data).

We see that light rail ridership is up 9%.  This appears low, given that the line and total investment were increased by about 33%, but the new extensions were not available for the whole year.  My guess correcting for opening dates is that on a full year basis this represents about a 20% increase.  For May, June, and July light rail ridership has been running 19%, 26%, and 20% above the same month last year (before either extension was opened).  This is well below the increase in line length (31%) and line total investment (36%)

I have always argued that the first 20 miles of the light rail line cut through the densest part of the city, including the downtown area (such that it is), the two highest visitation sports stadiums, and ASU  -- and as such any future expansions were going to have a much harder time justifying themselves (ie, as in this case, a 31% expansion in length would yield something less than 31% increase in ridership).  Light rail supporters have argued in return that I had things exactly backwards, that network effects would mean that ridership increased faster than route length.  My sense is that this argument will pretty clearly tip in my direction by the time we have 2017 data.

By the way, with total investment up to over $2 billion and average weekday round trip ridership at about 23,500 in fiscal 2016, then the total capital cost (not including annual operating subsidies) of the line sits at about $85,000 per round-trip rider.   Whenever Valley Metro supporters defend the line against my attacks, they will often quote various riders saying how much they love it.  Of course they do!  They damn well should love it -- we taxpayers spent $85,000 for each one of them to open the line AND subsidize every single one of their rides.

But if you really want to see the cost of these subsidies, look at the bus and total transit ridership in the chart above.   Total transit ridership in the area has fallen to the lowest point since before the light rail line was first opened, despite the fact the city it serves is still growing.  This is because bus ridership has fallen off the map.   Just last year, for every 1 rider gained to the light rail line expansions, 3.6 were lost on busses.   To see how far bus ridership has fallen, you have to go back further than the chart above shows.  Here is an older Valley Metro chart I annotated for a previous article:

You can see from this that bus ridership in Phoenix fell to a level we have not seen since 2003!  This is despite the fact that the Phoenix MSA has added about a million people since that time.

As it does in every city, light rail costs are starving the rest of the transit system.   By shifting transit dollars into a mode that requires 10x  more money to move a single passenger, the numbers of passengers served has to fall.  You can see from the chart that Phoenix transit ridership was rising steadily from 1997 to 2009.  But once light rail was completed, transit ridership absolutely stopped growing, despite population increases and large increases to total transit budgets.  Without light rail, we might very well have seen transit ridership as high as 90 million today, if past ridership growth trends held.

A two-fer!  This is from Yglesias's very good article on passenger rail also quoted in the previous post.  In discussing why Amtrak is generally uninterested in making incremental improvements on the Northeast Corridor, he writes:

The way Amtrak is currently set up, there's no real incentive to undertake incremental improvements. The Northeast Corridor already generates an operating profit, which simply defrays losses elsewhere in the system. Making it run better doesn't generate any wins for the people who would have to do the work, and would plausibly just lead Congress to reduce subsidies. If the NEC were spun off as an independent entity — perhaps even a private company — then it could internalize the gains from improved service and seek private financing to make cost-effective investments.

Long-time readers will know that my company privatizes the operation (but not the ownership) of public parks.  I will make two-hour presentations to parks agencies about how we can improve operations quality while cutting costs by 30-50% or more, and the near-universal response is, "well, if you reduce costs, then the legislature will just reduce our appropriations."  More efficient park operations, and at the margin better visitor service, don't create any wins for agency employees given their incentives.  In fact, if the parks are improved and more people show up, their job is just harder.  I had the manager of Arizona's premier state park tell me, absolutely in all seriousness, that he had the best job in the world if it wasn't for all the visitors.  Can you imagine a McDonald's franchise manager saying that?   As I have always said, government is not populated with bad people, it is populated with perfectly normal people who have terrible incentives.

When agencies choose how to spend incremental funds, they will almost always try to route these to the agency staff, in the form of more headcount and/or more pay.  When money is actually spent to make investments in the parks themselves, projects are chosen not by return on investment or customer priorities, but based on which ones will create the most prestige for the agency and its leaders.  This latter is one reason the Washington Metro is the mess it is, as the agency and the politicians who make appropriations will always prioritize system expansions over capital maintenance and sensible incremental improvements.

Yglesias has a very good article on why passenger rail is not a bigger deal in the US.   In it, he says this (emphasis added):

Instead the issue is that the dismal failure of US passenger rail is in large part the flip side of the success of US freight rail. America's railroads ship a dramatically larger share of total goods than their European peers. And this is no coincidence. Outside of the Northeast Corridor, the railroad infrastructure is generally owned by freight companies — Amtrak is just piggybacking on the spare capacity.

It is a short article, so it does not go into more depth than this, but I have actually gone further than this and argued that the US freight-dominated rail system is actually far greener and more sensible than the European passenger system.  As I wrote years ago at Forbes:

The US rail system, unlike nearly every other system in the world, was built (mostly) by private individuals with private capital.  It is operated privately, and runs without taxpayer subsidies.    And, it is by far the greatest rail system in the world.  It has by far the cheapest rates in the world (1/2 of China’s, 1/8 of Germany’s).  But here is the real key:  it is almost all freight.

As a percentage, far more freight moves in the US by rail (vs. truck) than almost any other country in the world.  Europe and Japan are not even close.  Specifically, about 40% of US freight moves by rail, vs. just 10% or so in Europe and less than 5% in Japan.   As a result, far more of European and Japanese freight jams up the highways in trucks than in the United States.  For example, the percentage of freight that hits the roads in Japan is nearly double that of the US.

You see, passenger rail is sexy and pretty and visible.  You can build grand stations and entertain visiting dignitaries on your high-speed trains.  This is why statist governments have invested so much in passenger rail — not to be more efficient, but to awe their citizens and foreign observers.

But there is little efficiency improvement in moving passengers by rail vs. other modes.   Most of the energy consumed goes into hauling not the passengers themselves, but the weight of increasingly plush rail cars.  Trains have to be really, really full all the time to make for a net energy savings for high-speed rail vs. cars or even planes, and they seldom are full.  I had a lovely trip on the high speed rail last summer between London and Paris and back through the Chunnel — especially nice because my son and I had the rail car entirely to ourselves both ways.

The real rail efficiency comes from moving freight.  As compared to passenger rail, more of the total energy budget is used moving the actual freight rather than the cars themselves.  Freight is far more efficient to move by rail than by road, but only the US moves a substantial amount of its freight by rail.    One reason for this is that freight and high-speed passenger traffic have a variety of problems sharing the same rails, so systems that are optimized for one tend to struggle serving the other.

Freight is boring and un-sexy.  Its not a government function in the US.  So intellectuals tend to ignore it, even though it is the far more important, from and energy and environmental standpoint, portion of transport to put on the rails. ....

I would argue that the US has the world’s largest commitment to rail where it really matters.  But that is what private actors do, make investments that actually make sense rather than just gain one prestige (anyone know the most recent company Warren Buffet has bought?)  The greens should be demanding that the world emulate us, rather than the other way around.  But the lure of shiny bullet trains and grand passenger concourses will always cause some intellectuals to swoon.

Which would you rather pounding down the highway, more people on vacation or more big trucks moving freight?  Without having made an explicit top-down choice at all, the US has taken the better approach.

Quote of the day from Randal O'Toole on the closure of the entire Washington Metro yesterday for emergency safety inspections

The Washington Post’s architecture critic claims that the shutdown happened because “we decided to let our cities decay.” In fact, it’s because politicians decided that spending money on new construction projects, such as the Silverand Purple lines, would benefit their political careers more than spending it maintaining the existing system.

Before that, it’s because politicians decided to saddle Washington with an expensive, obsolete technology that the region can’t afford to maintain. Metro needs to spend $1.1 billion a year on maintenance to keep the system from deteriorating; it spent about a third of that in 2014, so it’s getting worse every year.

In 2014, I published  an article based on Valley Metro's (our transit operator in Phoenix) very own ridership chart.  Here was the chart I showed:

My point was that the huge amount of money spent on light rail, which essentially constituted a single commuting route in this enormous and spread-out city, was cannibalizing bus service.   The cost and investment to carry a light rail passenger is at least an order of magnitude greater than that needed to carry a bus passenger, and no public system can long endure this sort of cost increase to shift passengers from a relatively cheap transit mode to an expensive one.  Inevitably, bus service (which mainly benefits the poor) is terminated to pay for the train (which benefits middle and upper class riders who would not be caught dead on a bus).  I am pretty sure the train would have been harder sell if they had been honest and said that they were going to have to cut bus service for the low-income working folks so that ASU students and Arizona Diamondbacks and Phoenix Suns fans could have a better way to get to the ballgame.

Anyway, Valley Metro has updated the chart for 2015 and it continues to look bad:

They were smart to cut off history on this one so you can't see how they killed the growth trend with the advent of light rail.  But you see that total transit ridership fell, with a small fall in light rail ridership and a huge fall in bus ridership.   Oops.  As an aside, they still have not fixed their terrible chart plotting.  You can see this years bar for light rail being longer than the one for 2013 despite the fact the number is lower.

The 2016 numbers will be interesting.  Since 2015 they will have opened several light rail extensions and got themselves a huge new tax increase approved (over my stern opposition).

Update:  In retrospect, the bus ridership fall was significant but "huge" probably is an exaggeration.  I was fooled by looking at the bar lengths, which again seem to have nothing to do with the actual data.  They are clearly drawing this manually -- no automatic charting program would get it this wrong.

Readers will know that this is one of my favorite topics on this blog, how huge investments in showy rail projects that amp up the prestige of government officials tend to cannibalize lower cost bus service and, at the end of the day, actually reduce total transit ridership.  The LA Times almost sortof recognizes this, and Randal O'Toole is on the case:

“Billions spent, but fewer people are using public transportation,” declares the Los Angeles Times. The headline might have been more accurate if it read, “Billions spent, so thereforefewer are using public transit,” as the billions were spent on the wrong things.

The L.A. Times article focuses on Los Angeles’ Metropolitan Transportation Authority (Metro), though the same story could be written for many other cities. In Los Angeles, ridership peaked in 1985, fell to 1995, then grew again, and now is falling again. Unmentioned in the story, 1985 is just before Los Angeles transit shifted emphasis from providing low-cost bus service to building expensive rail lines, while 1995 is just before an NAACP lawsuit led to a court order to restore bus service lost since 1985 for ten years.

...

Transit ridership is very sensitive to transit vehicle revenue miles. Metro’s predecessor, the Southern California Rapid Transit District, ran buses for 92.6 million revenue miles in 1985. By 1995, to help pay for rail cost overruns, this had fallen to 78.9 million. Thanks to the court order in the NAACP case, this climbed back up to 92.9 million in 2006. But after the court order lapsed, it declined to 75.7 million in 2014. The riders gained on the multi-billion-dollar rail lines don’t come close to making up for this loss in bus service.

...

Los Angeles ridership trends are not unusual: transit agencies building expensive rail infrastructure often can’t afford to keep running the buses that carry the bulk of their riders, so ridership declines.

• Ridership in Houston peaked at 102.5 million trips in 2006, falling to 85.9 million in 2014 thanks to cuts in bus service necessitated by the high cost of light rail;
• Despite huge job growth, Washington ridership peaked at 494.2 million in 2009 and has since fallen to 470.4 million due at least in part to Metro’s inability to maintain the rail lines;
• Atlanta ridership peaked at 170.0 million trips in 2000 and has since fallen nearly 20 percent to 137.5 million and per capita ridership has fallen by two thirds since 1985;
• San Francisco Bay Area ridership reached 490.9 million in 1982, but was only 457.0 million in 2014 as BART expansions forced cutbacks in bus service, a one-third decline in per capita ridership;
• Pittsburgh transit regularly carried more than 85 million riders per year in the 1980s but is now down to some 65 million;
• Austin transit carried 38 million riders in 2000, but after opening a rail line in 2010, ridership is now down to 34 million.

I will add that total transit ridership has been totally flat in Phoenix after construction of a major light rail project.  The project's total cost is approaching $2 billion as they slowly add on short extensions, but this amount did nothing but cannibalize bus ridership.  In fact, the situation is worse than this, since before light rail was built, Phoenix transit ridership was growing rapidly every single year, so in fact light rail actually likely reduced ridership by about 14 million.  The whole story is here.  (I will have an update in a moment but they have updated the chart from that article and ridership fell yet again in 2015).

This is one of those questions that seems like a no-brainer -- a bunch of people are sharing a ride, so they must be saving energy.  When asked this question, we all think of a full bus or train of people vs. the number of cars that would have carried the same people.

The key issue turns out to be occupancy -- how full is the train or bus.   And it turns out that occupancy is probably lower than most people think.  That is because everyone rides on buses or trains as they commute -- they are going in the direction of most people's travel at the time of day they travel, so the transit is totally full.  But no one thinks about those trains having to go back the other direction, usually mostly empty.   As a result, we get to this fact, from the National Transit Database as synthesized by Randal O'Toole.

2014 Energy Use per Passenger Mile

• Transit:  3141 BTU
• Driving:  3144 BTU

Valley Metro Rail here in Phoenix does better, at a reported 1885 BTU per passenger mile.   As reported many times here on this site, the cost of building this rail line, now well over one and a half billion dollars, would easily have bought every round trip rider a new Prius, with a lot of money left over.   This would have saved more energy as well.  Buses in Phoenix are averaging just over 6000 BTU per passenger mile.

A reader sent me this:  A recent McKinsey study looked at large infrastructure projects (over $1 billion).  In the introduction they observe:

Rail projects, for example, go over budget by an average of 44.7 percent, and their demand is overestimated by 51.4 percent.

We actually can combine these two numbers and find that the total cost per user of these systems was over-estimated by 117%, meaning the cost per user was on averagemore than double what was promised when these projects are sold to the taxpayers.

At the end of the day, the first segment of the line cost $75,000 per round trip daily raider.  We could have bought every regular rider a prius and a lifetime of gas and still saved money.

The AZ Republic has some of the first information I have ever seen on the nature of Phoenix light rail ridership.  The first part confirms what I have always said, that light rail's primary appeal is to middle and upper class whites who don't want to ride on the bus with the plebes

Light rail has changed the demographics of overall transit users since the system opened in 2008, according to Valley Metro.

Passengers report higher incomes than bus riders, with more than a quarter living in households making more than $50,000 a year. Many riders have cars they could use.

The 20-mile system running through Phoenix, Tempe and Mesa recorded more than than 14 million boardings last year. Still, census data estimate less than one-third of 1 percent of Phoenix commuters — or about 2,000 people — use rail as their main transportation to work.

.0033% huh?  If we built similar facilities to serve everyone, it would only cost us about $420 billion at the rate of $1.4 billion per third of a percent.

But I thought this next bit was the most startling.  I always had a sneaking suspicion this was true but never have seen it in print before:

While the much larger bus system reaches most corners of the Valley, light rail connects specific destinations along a single line. Nearly half of light-rail riders are enrolled in college.

I must have missed this in the original sales pitch for the light rail line: "Let's pay $1.4 billion so ASU students can get to more distant bars."   Note that by these numbers, students likely outnumber commuters 10:1.  Doesn't bode well for light rail extensions that don't plow right through the middle of the most populous college campus in the country.

Postscript:  They don't break out people riding to get to sporting events downtown, but sporting events make up most of the largest traffic days on the system.  From my personal acquaintances, many people use light rail as a substitute for expensive downtown parking at sporting events, parking (often semi-illegally) near light rail stops and taking the train the rest of the way in.  On the whole, its not very compelling as a taxpayer to be helping to subsidize someone else's parking.  And from a municipal fiscal standpoint, it means that light rail fares may be cannibalizing (on a much greater ratio than 1:1 given the price differential) parking fees at municipal parking lots.

I was reading this article in the Arizona Republic (which is generally an unskeptical cheerleader for light rail investments) and looking at the claim that Valley Metro (the operator of Phoenix light rail) had a 45% fare recovery ratio in 2013.  One would think that means that their fares cover 45% of their costs -- which would be awful for any real enterprise but is pretty good for government-run rail systems.

But in fact, Valley Metro (along with rail supporters in the Administration) stack the deck by defining most of the costs out of this metric so it looks far better than it really is.  In fact, by my reading of their financial statements, the true fare recovery is at best 15.2% and likely much worse.

Here are the Valley Metro financials for 2013, from their annual report

Look at their costs of $75+ million and fares of about $12.8 million.  How do they get to 45%?  Simple -- they leave the majority of their costs out.  They exclude administrative costs, financing costs, and depreciation (essentially their capital cost) from the equation.  This means that their fare recover is 45% -- IF you ignore their large administrative costs and you ignore the $1.4+ billion the line cost to build.   Further, because of the way the government does its finances, it is also missing any financing costs (interest and such on debt).

So the true cost recovery, stated in a way that  a reasonable person would think about such a number, is 16.8% at best.  If one takes into account the $8.28 million in "non-operating expenses" the number is 15.2%.  And it is even lower if you were to include interest and other financing costs.

I am sure Valley Metro will argue that this is the way the Feds measure it.  I don't care.  It has an obligation to accurately report its financial position to the public who is paying for it, particularly when the public is considering further investments in the form of higher taxes.  The 45% is a meaningless number that is crafted solely to make light rail look financial stronger than it really is.

This, by the way, is why total ridership of rail + buses has stalled since the construction of the light rail line.  Light rail is so expensive per rider that it is starving cash from the rest of the system.

Randal O'Toole and the Arizona Free Enterprise Club have weighed in with a comprehensive report on Phoenix's Prop 104 transit tax, and the results are ugly.  A few findings:

• The oft-repeated claim that light rail has generated $7 Billion dollars in economic development is simply untrue. In fact, many of the projects included in this claim have never been built (like the Sycamore Station development) or involve projects that have nothing to do with light rail (such as the $600 million Convention Center Expansion, which was funded largely by state tax dollars).
• The main beneficiaries of the transit plan appear to be contractors and developers who have projects near rail stations. The tax revenue from the plan combined with the generous subsidies offered to select developments ensures that this plan will benefit a few contractors and developers at the expense of others.
• The plan is unbalanced and ignores vehicle street improvements. Despite the fact that only 3% of the population uses transit (less than 1% use light rail), 95% of the funding in the plan goes toward expanded bus and rail service. Only 3% goes toward vehicle street improvements.
• Transit ridership actually fell after the light rail opened. From when light rail opened in 2009 through 2014, any gains in light rail ridership were offset by the loss of more than one bus rider. Ridership is still 1.2 million less per year than it was in 2009.
• The transit plan as proposed will increase traffic congestion, energy usage and greenhouse gas emissions. In fact, the transit plan will use more

The complete report is here.

Coyoteblog readers will be totally familiar with this statement from the report about light rail merely cannibalizing bus riders, echoing past articles I have had saying the same thing:

According to the city of Phoenix and Valley Metro, light rail is a great success in Phoenix, generating a 42-percent increase in transit ridership since 2001 and stimulating the construction of $7 billion in new real estate development along its route. A close look, however, reveals that both of these claims are wrong.

The increase in ridership took place between 2001 and 2009, the (fiscal) year that the light-rail line opened. Since that year, for every light-rail rider gained, the region’s transit systems lost more than one bus rider. Per capita transit ridership has declined by 8 percent since 2009 partly because the high cost of light rail forced a 34-percent increase in average bus fares by 2010 and an 18-percent decrease in bus service by 2013.

You can see this perfectly well from a chart right off of our transit authority's web site (except for my annotation in red), which I discussed in depth here.

It is just incredibly disingenuous that light rail supporters are trying to claim credit for transit ridership increases that occurred before the line was built and whose growth the line essentially halted.

Another claim the report demolishes is that light rail is somehow spurring development.  Supporters claim $7 billion of light rail development planned or built along the line, but oddly enough that is the exact same figure they were touting almost 8 years ago before the line was even completed.  Doesn't seem like they are getting much traction, huh?  In fact, the list actually shortens with every year as projects get cancelled and no new ones are added.  But beyond this, simply adding up development along the line and claiming that it is all incremental to the line's construction is simply moronic, the same facile BS analysis often used to support publicly funded football stadiums.    The obvious questions are:

• How do they know this development is incremental, and not development that would have occurred anyway?  In particular, this line was built through the three of the largest pre-existing development hubs in the metropolitan area (North Central Ave, Downtown Phoenix, and Tempe/ASU).  There was always development activity in these areas, and always going to be
• Much of the new activity they cite is near Tempe Town Lake, and I would give that project, not the rail, much of the credit.  The city did a marvelous job (see, I can give props to government once in a while) converting a big wide ugly dry ditch into a lake that is the centerpiece for business and condo development
• Much of this development is subsidized by various government programs.  It is impossible to separate the effects of the subsidy from the rail line.

Finally, if you don't believe me about the relative costs of the two modes, let's take a look at the number from Phoenix's own plan.  They speak for themselves:

So what does one get for the 5x higher operating costs and 134x higher capital costs of light rail over buses?  Well on the negative side, one gets a system that is substantially less flexible and responsive to changes.  The only positive I can come up with is that middle and upper class white people consider buses low class and want a transportation mode of their own.

A reader sent me a link to this Vox article on streetcars.  What I thought was interesting is just how weak the case for streetcars is, even when made by folks are are presumably sympathetic to them.  This page is entitled "Why do cities want streeetcars."  The arguments are:

• Tourists like them, because you can't get lost like you can on buses.  My response is, "so what."  Unless you are one of a very few unique cities, tourists are a trivial percentage of transit riders anyway.  Why build a huge system just to serve out-of-town visitors?  I would add that many of these same cities (e.g. Las Vegas) considering streetcars are the same ones banning Uber, which tourists REALLY love.
• Developers like them.  Ahh, now we are getting somewhere.  So they are corporate welfare?  But not so fast, they are not even very good corporate welfare.  Because most of the studies they cite are total BS, of the same quality as studies that say sports stadium construction spurs all sorts of business.  In fact, most cities have linked huge tax abatement and subsidy programs to their streetcars, such that the development you get with the subsidy and the streetcar is about what you would expect from the subsidies alone.  Reminds me of the old joke that mimicked cereal commercials: "As part of a breakfast with juice, toast, and milk, Trix cereal has all the nutrition of juice, toast, and milk."
• Good for the environment.  But even Vox asks, "as compared to what."  Since they are generally an alternative buses, as compared to buses that have little environmental advantage and often are worse (they have a lot more weight to drag around when empty).
• The Obama Administration likes them.  LOL, that's a recommendation?  When you read the text, what they actually say is that mayors like the fact that the Obama Administration likes them, for it means the Feds will throw lots of Federal money at these projects to help mayors look good using other peoples' money
• Jobs.  This is hilarious Keynesianism, trying to make the fact that streetcars are 10-100x more expensive than buses some sort of positive.  Because they are more inefficient, they employ more people!  One could make the exact same argument for banning mechanical harvesters and going back to scythes.   Left unquestioned, as Bastiat would tell us, is how many people that money would have employed if it had not been seized by the government for streetcar use.
• Je ne sais quoi.  I kid you not, that is their final argument, that streetcars add that special something to a neighborhood.  In my mind, this is Vox's way of saying the same thing I did the other day -- that the streetcar's appeal is primarily based on class, in that middle and upper class folks don't want to ride on a bus with the masses.   The streetcar feels more upscale than buses.   The poor of course, for whom public transit is most vital, don't want to pay 10 times more for sexiness.  Oh, and watch this video of Washington streetcars blocking traffic and crunching parked cars and tell me what it is adding to the neighborhood.

Every argument I have ever been in on streetcars always boils down to something like "well, all the cool kids like them."  Once, after defending the US approach to rail (vs. Europe and Japan) as (correctly) focusing on productivity vs. sexiness, having gone into a lot of detail on the economics of freight vs. passengers, I got a one sentence answer from Joel Epstein of the HuffPo:  “You should get out of the country more often.”  That was it -- the cool cosmopolitan kids who vacation in Gstaad but never would be caught dead driving across Nebraska were all against me.

I am increasingly convinced that the appeal of streetcars and light rail has everything to do with class.  From any rational perspective, these systems make no sense -- they are 10-100x more expensive than buses and lack the flexibility that buses have to adjust to shifting demand patterns over time.  A single extra lane of highway adds more capacity than any light rail line.

Streetcar's single, solitary advantage is that middle and upper class whites who would not be caught dead on a bus seem to be willing to ride streetcars.   I don't know if this is because of some feature of the streetcars (they are shiny and painted pretty) or if it is some sort of self-segregation (the upper classes want to ride on something that is not filled with "riffraff").

Anyway, Reason has a really good video on streetcars, which covers a lot of topics from cronyism to the class issues I mention.  And the film of the incredibly poor design of the DC streetcar has to be seen to be believed.  (via Maggies Farm and their great daily links page)