Coyote Blog

Patrick Everson has the first of a two-part series of editorials in the Las Vegas Review-Journal on light rail.  In this first part, he is nice enough to refer to much of what I have written here about the problems with Phoenix Light Rail.  You can find his editorial here.

I bring your attention back to this chart from this post the other day about light rail killing transit growth.

I have no evidence that this chart was deliberately manipulated, but somehow the light rail ridership bar for 2014 got exaggerated.  It certainly seems suspicious.  Light rail ridership went up from 2013 to 2014 by only about 45,000, or 0.3%.  This is negligible  We should not even see the bar move.  Note the total ridership in 2011 and 2010 when ridership fell by 86,000 but the bar lengths are almost indistinguishable.  The rail ridership looks to my eye like the bar is 7-9% longer, not 0.3% longer.  In fact, the bar for 2014 clearly goes past the halfway point between 10 and 20, despite the fact that 14.3 should be less than halfway.  In fact, the 2014 rail increase of 45,000 is graphed as visually larger than the 1.3 million decrease in busses.

Check this graph out from the Phoenix Metro web site.  It shows bus ridership in years past, and more recently both bus and light rail ridership.

You can see a few things.  First, note that almost all the rail ridership came at the expense of bus ridership.  It  was almost a pure 1:1 substitution.  The bus ridership, even with a half year of light rail being open, was 65.7 million in 2009.  Total ridership was only 67.6 million in 2010 and 2011.  Yes there is a recession here, but of the 12 million or so in light rail ridership, at least 10-11 million of that came out of buses.  Essentially, we paid $1.4 billion in capital costs to move 10 million riders to a mode of transit that is at least an order of magnitude more expense.  Nice work.

Second, note that after over 12 years of growth, with the onset of light rail transit ridership has stagnated for 6 years.  Some of this, at least initially, is likely due to the recession but in fact recessions are supposed to spur transit ridership, not reduce it, as people look for lower cost alternatives.  There is a good explanation for this.  Because light rail is so much more expensive, the cost per rider for the entire transit system has skyrocketed.  With budgets unable to be increased this fast (and with fares covering only a tiny percentage of rail costs), the system must cut back somewhere.  Since rail can't really be cut back, bus routes are cut.

If we had seen the same growth rate from 2009 to 2014 as we had seen in the twelve years prior, we should have over 86 million trips in 2014 (note these are fiscal years, and fiscal year 2014 is already closed, so this is not partial year data).

We paid, and continue to pay (since rail must be subsidized heavily) billions of dollars to reduce transit ridership.

The Tampa Rail blog has responded to my post criticizing Phoenix light rail (which the Tampa folks used as a glowing example of rail success).  Remember I wrote, in part:

Look, I don't think I have ever argued that Phoenix Light Rail was run poorly or didn't have pretty trains.   And I don't know if moving 18,000 round trip riders a day in a metropolitan area of 4.3 million people is a lot or a little (though 0.4% looks small to me, that is probably just my "pre-web" thinking, whatever the hell that is).The problem is that it is freaking expensive, so it is a beautiful toy as long as one is not paying for it.  Specifically, it's capital costs are $75,000 per daily round trip rider, and every proposed addition is slated to be worse on this metric (meaning the law of diminishing returns dominates network effects, which is not surprising in this least dense of all American cities).

Already, like in Portland and San Francisco, the inflexibility of servicing this capital cost (it never goes away, even in recessions) is causing the city to give up bus service, the exact effect that caused rail to reduce rather than increase transit's total share of commuters in that wet dream of all rail planners, Portland.  Soon, we will have figures for net operating loss and energy use, but expect them to be disappointing, as they have in every other city (and early returns were that fares were covering less than 25% of operating costs).

Of course, as with all government issues, the ultimate argument is that I am some sort of Luddite for actually demanding definable results for billion dollar government spending

Sorry Coyote, save for the topic matter I'm afraid I'm just not going to be much fodder for you. We're years past 'it's an expensive tax thing'.

We know that. We know rail like any capital project is expensive to execute and expensive to maintain - in dollars. But anyone who raises the math to me will wind up with the same big 'so what'. Community investment doesn't bother everyone the same way and different people see different value. There's no way you or I cold supernaturally understand the net benefit for or against light rail. We must simply choose to believe and pick our sides.

If you believe that just because rail is expensive they aren't worthwhile, you need to explain every public vote that has gone for implementing and expanding rail systems around the world even though most operations are publicly subsidized.

Gotta run'em well, and, over time, integrate with a city, but LR is a carefree mobility solution in areas where people choose to support and pay for it.

See, they are well past my neolithic argument, into their little post-modernist world where aesthetics and political correctness trump any actual need to demonstrate money is being used well.  Though it is interesting to see him resorting to faith as a justification.

I have two words for this person -- "opportunity cost."  On one hand, the money for this project must be taken out of private hands to build the rail line -- even leaving out the substantial individual liberties questions here, there is still some obligation to demonstrate the money is better used than it would have been in the private hands from which it is taken.  Ditto, by the way, for the stimulus bill.   On the other hand, to the extent that one wishes to spend government money to move people from A to B, one needs to demonstrate that this method is better than others.  I would argue high speed rail fails both tests.

Update: Joel Epstein and I have a go around the same issues in the Huffpo comments.

Tampa Rail writes (hat tip to a reader):

The new Phoenix light rail system is emerging as one of the most successful new systems in the country.  This is especially poignant for Tampa because in scale, project scope, and demographics, Phoenix represents the apogee of operating examples.

Over the course of its first year the system has received high marks in community integration, stunning ridership figures, and respectful financial constraint (making tough decisions on long-term planning that do not inhibit the value of its starter-line status today).  This is exactly what Hillsborough County is shooting for in its own implementation.  A perfect balance of conservative control and benchmarking combined with progressive action and democratic freedom, the latter which may finally come to Hillsborough County in the form of a referendum.  That all good stuff was achieved by such a strikingly similar auto-depenent culture is a great omen.  A starter light rail system can be championed by civic conservatives (Mark Sharpe), and civic progressives (Ed Turanchik) to great outcome....

Both pieces I link to here embarass anti-rail or anti-tax groups who are, as the Phoenix article notes, "muted" if not definetively silenced.  Their arguments against community investement were loud, often intelligent (once one bought into the ideological premise that rail systems must 'pay for themselves' and that community investement is somehow inherently evil - points not firmly established by any means among rationale individuals), and grossly atypical.  I will forever hype on how mechanical, unchanging, and how pre-web these attacks were formulated.

Ooh, how can I overcome my embarrassment?  Look, I don't think I have ever argued that Phoenix Light Rail was run poorly or didn't have pretty trains.   And I don't know if moving 18,000 round trip riders a day in a metropolitan area of 4.3 million people is a lot or a little (though 0.4% looks small to me, that is probably just my "pre-web" thinking, whatever the hell that is).

The problem is that it is freaking expensive, so it is a beautiful toy as long as one is not paying for it.  Specifically, it's capital costs are $75,000 per daily round trip rider, and every proposed addition is slated to be worse on this metric (meaning the law of diminishing returns dominates network effects, which is not surprising in this least dense of all American cities).

Already, like in Portland and San Francisco, the inflexibility of servicing this capital cost (it never goes away, even in recessions) is causing the city to give up bus service, the exact effect that caused rail to reduce rather than increase transit's total share of commuters in that wet dream of all rail planners, Portland.  Soon, we will have figures for net operating loss and energy use, but expect them to be disappointing, as they have in every other city (and early returns were that fares were covering less than 25% of operating costs).

PS- I get a lot of comments that I have some weird anti-train bias.  Actually, I have an n-scale model railroad in one room of my house, and spent a lot of my teenage years traveling along rural rail lines and photographing trains.  I love trains.  I just don't like stupid investments.

PPS- I was just thinking, on the basis the Tampa writer declares the building of Phoenix Light Rail a raving success, I could say the same thing about buying a super-size 100" flat screen TV for $50,000.  It is beautiful.  Everyone who sees it will love it.  It works flawlessly.  Lots of people will be able to enjoy it at one time.  In fact, it is the greatest, most sensible and successful purchase of all time as long as you never mention the cost.  Which is, by the way, why only one person I have ever met has one (I happened to be at a Reason reception the other night and the homeowner had such a beauty on his living room wall).

Update: I try to anticipate every argument in these posts.  The one other argument is that rails takes congestion off roads.  But for most of its length, Phoenix light rail displaced one lane of road in each direction.  These lanes had a capacity as large or larger than what Phoenix light rail carries.  The were also much cheaper to build.  I must say I liked my quote from that post

If running trains requires, as you suggest, draining resources from millions of people just to move thousands, how is it sustainable?

I wondered today what kind of subsidy a rider on the Phoenix Light Rail system was receiving.  Hillary Foose, the public information officer of Metro light rail, was kind enough to send me a link to this board presentation.   Since the rail system opened mid-fiscal year, I will use their own projections for the 2009/2010 fiscal year.

Public accounting is a pain in the butt for someone used to private finances, because it is all cash accounting rather than accrual and they mix together capital expenditures with operating expenditures.  But the table on page 62 carves out the operating budget for the existing 20-mile line from the development and capital budgets.    Here are the key numbers:

Fare Revenue:  $8,985,159

Operating Expenses:  $33,733,168

So already on an operating basis we have a 73% subsidy.  But we have sunk $1.4 billion of capital money into building the line  (actually this is a little low as Metro has spent tens of millions more this year).  Unfortunately, in government accounting, there is no depreciation or interest charge that shows up.   So I am going to charge them with the payment on a 30-year $1.4 billion 5% note, which would be just over $91 million a year.

Totaling the $91 million with the other operating expenses, we get a 93% subsidy for light rail.  This means the true cost of the $1.75 ticket for a light rail ride is actually $25!  METRO says that light rail riders love the service.  I should think anyone who gets a $25 service for $1.75 should be happy.

Another way to look at the subsidy is on a per rider basis.  So far, METRO has averaged about 17,000 round trip riders per weekday (based on about 34,000 boardings per day).   The $115.8 million annual subsidy (capital+expense minus revenues) works out to just over $6,800 per rider per year that the rest of us (who may not live or work near the line**) pay each current rider.

There are a number of ways in which I have likely understated the subsidy:

1. I used their June revenue projections, which likely will continue to be revised downwards as ridership continues to slump
2. I used their own expense projections, and we know how often governmental bodies hit their expense numbers
3. I assumed no new capital spending necessary over a 30 year life.  Rail experience has shown this to be overly-optimistic.  Rail lines have to be rebuilt every 15-20 years or so.  They take tons of capital maintenance dollars.   When we look back twenty years from now, we'll likely come to the conclusion I grossly understated the capital charge.

**Footnote: Since over a third of the capital to build the line came from the Feds, many of the people subsidizing the METRO riders don't even live in this state.

Update: The other thing I left out is lost parking revenue.  The revenue numbers for fares is in fact overstated.  It should net out lost parking revenues, for example at baseball games.  This is the only time I ride the Metro, because I substitute a $2.75 Metro round trip ticket for a $10 city garage parking expense.  But the city has never acknowledged this cannibalization.

Update #2: I have posted an update here

I am always amazed at the lengths to which some folks will try to put lipstick on the light rail pig.  One example I found today.  Michael Graham Richard wrote on treehugger in June:

The sprawling city of Phoenix, of all places, is showing us how light rail should be done. They just opened a 20 mile line with 28 stops last December, and ridership statistics are beating all forecasts (evidence that the same might be true in other cities where they are afraid to invest because their forecasts are too low) with 40,000 daily riders instead of the 25,000 expected.

But here are the ridership figures from Valley Metro, who runs Phoenix Light Rail.  This is weekday ridership (actually number of daily boardings) -- weekend ridership is much less:

• Jan:  30,617
• Feb:  35,277
• Mar:  34,376
• Apr:  37,386
• May:  33,553
• Jun:   29,469
• Jul:  26,554

It is hard to see where one gets a 40,000 figure, especially since a true daily rider/boarding figure would have to average in the lower Saturday/Sunday numbers.

And who cares if it meets some sandbagged forecast or not?  Is 40,000 even a reasonable number?  Note that even at the higher 40,000 figure this implies just 20,000 round trip customers.  This higher ridership number would still make the capital cost of the $1.4 billion line to be $70,000 per round trip rider, and ABSURD subsidy.

Update: The ridership numbers will likely pick up when Arizona State is back in school.  ASU and the baseball stadium are about the only major destinations on the line through dispersed, low-density Phoenix (it goes through our "downtown" but that is not saying much  -- it is not a big center of employment).  Did we really build light rail as another subsidy for ASU students?

Update #2: Let's say there are 50,000,000 big city commuters in the US in cities outside of Boston/NY/Chicago with large transit systems.   Serving these commuters at $70,000 each would create a capital cost of $3.5 trillion for light rail.   Who on the planet really thinks this is reasonable?  Sure, you would get some network effects as you built out lines that increased ridership, but these would be offset by diminishing returns (presumably the first Phoenix line was built on the most promising corridor, and all future corridors will be less promising).

The Arizona Republic had another of its cheerleading articles on light rail this morning.  In it was a chart that, contrary to the intent of the article, summarized exactly why Phoenix light rail is doomed.  Below is a chart of the employment density (top chart) and population density (bottom chart) at each stop along the first rail route.  Note that this line goes through what passes for the central business district of Phoenix and the oldest parts of town, so it was chosen to run through the highest density areas - all future extensions will likely have lower numbers.  Unfortunately, they do not reproduce this chart online so here is a scan:

Take the population density chart.  As a benchmark, lets take Boston.  The average density for all of the city of Boston is 12,199 people per square mile.  Phoenix's light rail line cut through the highest density areas of town has only one stop where density reaches this level, and most stops are less than half this density.  And this is against Boston's average, not against the density along its rail routes which are likely much higher than the average.

Rail makes zero sense in a city like Phoenix.  All this will do is create a financial black hole into which we shift all of our bus money, so the city will inevitably end up with a worse transportation system, not a better one.  Cities that build light rail almost always experience a reduction in total transit use (even the great God of planners Portland) for just this reason - budgets are limited, so since rail costs so much more per passenger, other transit is cut back.   But the pictures of the train will look pretty in the visitor's guide.

Postscript: Phoenix's overall average density is around 2,500 per square mile.  Assuming that the 12,000 in the chart above is one of the densest areas of Phoenix, this gives a ratio of about 5:1 between peak and average density.  This same ratio in Boston would imply peak density areas of 60,000 per square mile.  This may be high, but indicates how much higher route densities on Boston rail should be.  Oh, and by the way, Boston rail is losing a ton of money.

Other city densities here from 1990.  People think of LA as spread out, but LA has a density over three times higher than Phoenix!

Tom Kirkendall observes that this could have been written about Houston light rail.  I would add that it also could have easily been written about Phoenix light rail, which I have criticized here and here and here.  And heavy rail? Don't get me started.

Beyond these impressions, Tom Rubin observes that VTA has "the worst
operating statistics fo any American transit operator." The reason for
this, he says, is that San Jose "” being built mostly after World War II
"” is one of the most spread-out urban areas in the country. Not only
are people spread out, but jobs are spread out, with no job
concentrations anywhere.

This makes large buses particularly unsuitable for transit because
there is no place where large numbers of people want to go. So what was
VTA's solution when its bus numbers were low relative to other transit
agencies? Build light rail "” in other words, use an expensive
technology that requires even more job concentrations.

Now it has one of the, if not the, poorest-patronized light-rail
systems in America. So what is its solution? Build heavy rail, a
technology that requires even more job concentrations.

This is an interesting factoid from another Anti-Planner post:

The amazing thing to the Antiplanner is that anyone would take this
proposal seriously. The average urban freeway lane costs about $10
million per mile. The average light-rail line costs about $50 million
per mile and carries only a fifth as many people. Seattle's proposed
lines were going to cost $250 million per mile, making then 125 times
more expensive at moving people than a freeway lane.