NARPAC RUMMAGES THROUGH THE WMATA BUDGET
This organization has pored through DC's city budget, its school budget, and those of many of the other local jurisdictions in our metro area. Now we have poked around in the approved 2004 budget of the Washington Metro Area Transit Authority. We are not searching for petty scandals or inefficiencies in the spending of public funds. We are trying to understand just what it takes to run a major metro area public transit system involving almost 340 million riders on 1500 buses on 161 routes (with 316 variants) over thousands of street miles, plus 1000 heavy rail subway cars running on 100+ miles of track, mostly underground, on six color-coded routes. The following generalities stand out.
o Bus routes are substantially lower density and more diverse, with no provisions for riders' cars. Rail routes are connected to bus routes and (often) parking lots. Single buses are "retail"; trains with 4-8 cars are "wholesale". Both systems have shown modest 5-year growth
o Both systems are almost equally manpower-intensive (per rider) butthe functional breakouts are distinctively different. The bus system puts almost all its effort into its vehicles, while the rail system puts half of its efforts into its privately owned rights of way, including, of course, elaborate stations.
o personnel costs account for about 70% of all operating costs, and the work force is highly unionized. Average salaries and wages are almost identical to those of the DC government (a readily available public sector budget for comparison), but WMATA fringe benefit costs are significantly higher. Three- quarters of total direct pay is in hourly wages, one-quarter in salaries. Overtime is surprisingly low, and dropping.
o Rail revenues pay over 70% of their system operating costs, bus revenues cover barely 30%. The remainder is derived from a (probably overly) complex set of jurisdictional "subsidies" wherein those entities that use public transit the most, must levy the highest bite on their taxpayers. A perverse NARPAC subsidy system would levy the greatest contributions from those jurisdictions with the most single-occupant car commuters. It would save DC about $125M per year, and cost Fairfax County about $90M.
o There are very large variations in bus route density and rail station density, making generalities difficult at best, and thoughts about savings due to route consolidation or station closings very unproductive.
o While the 2004 capital budget has quite specific plans in its three major categories of "infrastructure renewal"; "system access/capacity growth" within the existing system; and "system expansion"; the amounts shown for the next five years are a formula for disaster. NARPAC believes WMATA should be spending $800M on system expansion, and $700M on maintenance and upgrades annually.
Ridership on DC's Metro system has been growing slowly but steadily over the past five years. In 2000, 138.5 million individuals took the bus, and 163.3 million took the subway. By June 2004, bus had grown 1.9% a year to 149.6M, and rail grew 3.7% yearly to 189.2M. As shown on the chart to the right, transit vehicles grew a bit faster. By mid-2004, there were nominally 1450 buses operating separately on hundreds of different routes, while 1000 rail cars operate on six routes in trains of 4 to eight cars. Revenue miles have increased somewhat faster: bus travel growing 4.3% yearly to 50 million miles, and rail 6.3% yearly to 63.5 million miles.
To run both systems, expenditures (upper right, below) have grown from about $640M in 2000, to a planned $872M in 2004, an annual growth rate of almost 19%. Rail operations will require about $500M in 2004, bus operations, about $333M, while "Metro Access" operations (the virtually "free" transport of the handicapped by individual vehicles) will reach $40M, doubling its costs in five years.
Revenues (upper left below) come partly from the passengers themselves,
and small amounts from advertising, bus charters, or metrorail parking.
Most surprising to many, 47% of total expenditures are recovered from
"subsidies" paid by the various jurisdictions (including DC) which use
the services. This is discussed in greater detail further along in this
Bus and Rail revenues are compared in the two lower charts immediately above, showing that the bus system only recovers about one-third of its expenses directly, while rail pays a good 70% of its expenses. Parking also offers a somewhat more remunerative add-on than advertising or any other secondary source.
As is typical of service operations, roughly three-quarters of all WMATA costs are directly for personnel, either in straight wages and salaries and overtime, or in the ever-growing category of fringe benefits (below left). Hourly wage jobs consume over 70% of regular pay, with salaried jobs taking the rest. There is a slow shift towards more salaried personnel and less wage earners. On the other hand, overtime is a very small share of the total, and has been declining. Fringe benefits are also growing slightly, making up 22% in 2004, compared to 20% in 2000.
Typical average pay scales (below left) may seem somewhat high, but in fact most of the jobs are skilled, and much of the technology moderately high. Train operators earn somewhat more than bus drivers, but rail and bus mechanics make very close to the same amount. Salaried workers make somewhat more than hourly wage workers, but not very much. The average wage earner makes $50, 675 in take-home pay, while the average salaried worker takes home $56,123.
The average pay for the all 10,000 WMATA employees comes out to $52,092, and as a point of comparison, that compares to $52,917 for the average of the 26,000+ employees in the DC government (blue bar). Oddly enough, it is the difference in fringe benefits that makes WMATA jobs the more desirable. The average fringe benefits for DC government workers is about $7,700 while for WMATA workers it is almost twice as much: $14,400. On the other hand, Montgomery County runs a substantial local bus network called "Ride On" with more than 80 routes connecting various county locales as well and Metrobus and Metrorail nodes. With over 200 buses and about 600 FTEs, its average pay is only $43,383, although its average fringe benefits reach $13,700 (green bar at bottom).
NARPAC cannot really explain Ride On's lower pay scale, although it is a substantially smaller and less complex operation than WMATA. More comparisons would be required before jumping to the conclusion that WMATA is somehow cost ineffective.
It is perhaps worth pointing out (above right) that the WMATA work force is quite highly unionized. All of the hourly wage workers, and 44% of the salaried workers are represented by one of several unions. NARPAC cannot draw any educated connection between the unions and the fringe benefits.
It is of considerable interest to NARPAC to look a bit more closely at the functional breakout of the 4085 workers in the bus system and the 5004 workers in the rail system. The largest two categories of workers are vehicle operations and vehicle maintenance, as shown in the left hand chart below. There are a little more than twice as many manhours in operating 1450 million- dollar buses than in operating perhaps 200 trains. However, maintaining 1000 $5 million-dollar rail-cars requires almost as many personnel as maintaining the larger numbers of buses.
On the other hand, Metrorail has several categories of personnel requirements missing from the bus world. These involve maintenance of the escalators and elevators required to reach the rail platforms, the more complex systems that support those trains, and the very large tasks of maintaining the tracks, roadbed, tunnels, and platforms. To this must be added the 400-odd special Metro police that patrol the rail system, compared to less than 100 related to the bus system.
As in the age old comparisons between the operating costs of freight railroads and trucks on the highway, the greatest difference is in the requirement that the rail systems own, maintain, and guard their own dedicated rights of way, while trucks get a free ride on highways used by others, and supported by other government agencies such as the Federal Highway Administration. Hence the charts below shown the right of way (RoW) costs "below the line" and the vehicle-related costs above. If we had also allocated the management and indirect support categories between vehicles and RoWs, we would find the Metrorail is almost 50/50.
As an aside, it is interesting (to some) that while the Metrorail trains spend almost exactly twice as much for electric power ($30M) as the buses do for diesel fuel and natural gas (while carrying more riders further and faster), the extensive network of tunnels and (often air-conditioned) stations, obliges Metrorail to spend almost five times as mush ($20M) on primarily electric utilities. No free street lights and atmosphere control for the subways!
One of the most extraordinary facets of WMATA operations is the complicated formulae worked out with DC and its neighboring jurisdictions to decide who pays how much of the "subsidy" required to make up for the fact that the systems do not earn enough to pay their own way. Without exploring all the ramifications, the total shortfall in revenues, direct or indirect, is allocated in the following proportions:
for bus shortfalls: (differentiate regional v non-regional routes, apply these weightings
to the regional routes)
for rail shortfalls:
In addition to virtually all of the costs of the Metro Access system ($39M), it should also be noted that the capital investment (right of way) costs for the Metrorail system are still being paid off through principal and interest payments on 30-year bonds. These are a part of the total shortfall, but should be allocated, of course, entirely to the Metrorail side. This debt service has stayed relatively constant at $27.5M over the past five years, with $16.5M in principal repayment, and $11M in interest. But together they comprise some 16% of the total shortfall (i.e., subsidy needed) of $408M in 2004.
What strikes NARPAC as odd is that these allocations are primarily in direct proportion to how much each jurisdiction uses these systems rather than how little. In short, don't use public transportation systems, and don't contribute (much) to the needed subsidy!
The lefthand chart below shows the composition of the 2004 subsidies to be paid by the six county/city jurisdictions (lumping Fairfax City and Falls Church under "Fairfax"). DC, which is the heaviest user of public transit (and has the most metro stations!) will pay almost $160M in 2004 (from its local operating budget!), while Fairfax, by far the richest county, will pay only about 35% as much ($56M).
NARPAC, going where political angels would fear to tread, has posited two simplistic alternate subsidy allocations. These indicate the range of differences that might be realized, and are shown in the right graphic above. In both cases, all the weighting factors are eliminated, and a single jurisdiction-peculiar parameter is put in its place. In the first case, each jurisdiction pays in proportion to its calculated ridership. The middle bars below show that DC and Fairfax would pay somewhat less, and the others (save Pr. George's County) a bit more. The major savings of just over $15B would accrue to DC.
The right hand (pale blue bars) show what happens if the shortfall is allocated in proportion to those who don't use mass transit, but commute alone by car to work(data which are available from Census data). In essence, the taxpayers in these jurisdictions kick in for not using the Metro system, which seems eminently sensible to encourage "smart growth". In this alternative the benefit to DC is very substantial ($125M per year!) While Virginia's Fairfax County's payment more than doubles (but not to as much as DC is paying now), the Maryland counties of Montgomery and Pr. George's pay less than 20% more. It makes perfectly good sense to NARPAC.
Bus operations are far more complex and difficult to analyze than rail operations, primarily because of the extraordinary numbers of different routes and "sub-routes" (NARPAC term, not WMATA's) involved. For instance, in DC there are 57 different "line names" (like "Sibley Hospital/Stadium Armory" and "Minnesota Ave/Anacostia") but a total of 100 variants of those routes. Second, there are a total of 32 'regional routes" like the two examples above) and there are 22 "non-regional routes" which make smaller loops such as "Capitol Heights/Benning Heights", or "Glover Park/DuPont Circle." Further, there is no equivalent term for "passenger miles". Instead, the basic parameter is "platform hours" which is an obscure way of defining how many hours the bus (and its driver, presumably) are away from its home garage (one of ten in the system).
In any event, in 2004, the 1450 buses are planned to be away from their garages for over 3.9 million hours, and travel almost exactly 50 million miles. This results in an average "fleet speed" (again a NARPAC term) of almost 13 miles per hour, thus racking up about 34,500 miles a year per bus over about 2700 hours per bus, or on the order of 10 hours/day, a very substantial utilization rate.
Metrobus expects to carry just short of 160 million passengers but collect from them only $97.3 million, or 61 cents per passenger. This amounts to only half the normal fare of $1.25 "due to transfers and flash pass impact" (according to their footnote). Because operating costs are expected to exceed $333.2M, the operating subsidy discussed above will reach $217M, or $1.36 per passenger!
There is, of course, great variety in the utilization of buses by different jurisdictions, and some variety in the proportion of regional to non-regional route structure. The left chart below shows that DC routes use almost 50% of all platform hours, with Maryland and Virginia split 30/20. 22% of all platform hours is for non-regional routes, and this varies somewhat between major jurisdictions. The right hand chart simply shows the great range in route "density" for all jurisdictions, and the greater density of some major regional routes. DC's big winner is the combination of the five sub-routes on the Pennsylvania Ave Line (#30, 32, 34, 35, 36) with 159,000 platform hours in 2004. Maryland service is heaviest on the C2 and C4 buses of the Greenbelt-Twinbrook Line with 82,000 hours, and Virginia's most heavily traveled bus route is along Columbia Pike with almost 55,000 platform hours split among several sub-routes.
It is of some concern that only one of the ten most heavily utilized bus routes in DC runs its length East of the Anacostia (in bold) where many of DC's most needy and car-less residents live, and one other barely crosses the River on Benning Road to the Minnesota Avenue Metro Station:
By comparison, the relatively "wholesale" movement of passengers by Metrorail along the six color-coded routes that start and end in the suburbs (but intersect in downtown DC) seems relatively simple to program and analyze. This first-class, relatively new subway system will carry 189.2 million passengers in 2004 using just about 1000 rail cars in 4 to 8-car trains for a total of 65 million railcar miles. Total railcar hours of operation are not provided in the budget statement, but the average speed must be considerably higher to rack up 65,000 miles per railcar per year. Total Metrorail costs, as discussed above, will run to almost exactly $500M, but passenger revenues will provide a good two-thirds of that ($330M). Hence, along with other small income sources, the subsidy required will be less than $131M, equating to 69 cents per passenger.
While the available figures on bus utilization do not readily indicate how many riders commuted to DC from anywhere in the suburbs, the numbers of riders entering and exiting from each Metro station is available, and have been used by NARPAC for other purposes. The chart below shows how the rush hour influx varies at each of the entry gates in June 2004. By far the biggest daily onslaught comes from Northern Virginia with more than 41,000 rush hour commuters, while the others vary from 10,600 to 14,900. This presents some difficulties in railcar utilization rates, since the train's length is not varied as it shuttles from one end of the line to the other. The biggest problem is on the Blue/Orange Line which needs over 400 railcars per rush hour from the West, but less than 130 from the East. How Metro solves this imbalance, NARPAC frankly does not know. However, it would certainly make sense to run many trains from Northern Virginia only as far as downtown and back.
Growth of 3600 Red Line riders between 1987 and 2004 (18 years) has averaged about 1.8% per year, and is by no means evenly distributed between the stations of the line, as shown on the chart below. Riders have doubled at the Shady Grove end of the line (15 miles from downtown), but very little at Glenmont at the other end. In between, the increases are much smaller, and in a few cases have dropped somewhat. Downtown increases (from DuPont Circle to Union Station) have been large, since they are the destinations for those arriving for all the other stations. The very large growth at two stations results from added activity at the new MCI Center at Gallery Place, and new local and federal office buildings near Union Station.
The next sets of graphics are an attempt to depict how riders on the morning rush hour trains build up from the outer extremities of the routes and then rapidly drop off in the downtown area as they reach their destinations. The left hand chart below shows activity on the Red Line with red bars (above the line) showing the entries during the peak hours at each station, and the green showing the exits (below the line). The paler colored bars show the lower activity in off-peak hours. The solid lines above are the integrals of the entries and exits during rush hour. This reinforces the notion that almost half the total rush hour riders from the Shady Grove line get on at that station, while the entries are more evenly divided along the route from Glenmont.
The smaller chart on the right above shows the same trends for the Green Line rush hours only, and the purple integral line indicates the portion of the total activity within DC. (It doesn't show as well on the left hand chart). Again however, the DC commuters are but a small share of those coming in from the suburbs to their jobs in the city. Both charts also indicate a significant difference in the extent to which each metro station is used. This provides a useful indicator, no doubt, of the relative number of residents with jobs elsewhere to commuters with jobs near each station. For instance, there are more working residents than jobs in Anacostia, but more jobs than residents around the Navy Yard.
Lastly NARPAC has tried to show the same build-ups for the Yellow, Orange, and Blue Lines reaching out into the relatively prosperous and densely developed suburbs of Northern Virginia to the left, and to the less developed suburbs of Southern Maryland. The "jump-ups" in the integral lines are where the subway lines merge It should also be noted that "downtown" starts as soon as the Blue/Orange Line enters DC from Virginia, since Arlington and Alexandria are "inner suburbs" (once part of DC), equivalent to those in DC's NW, NE, and SE.. This further shows the imbalance between the West and East influx to the city, and the fact that the Rosslyn "portal" to the city is twice to three times as crowded as the other approaches, and is, in fact, on the verge of saturating this part of the system..
After going for some seven years without fare increases, Metro has been obliged to raise fares twice within the span of a year. Officials seemed somewhat surprised that it has had no negative effect on ridership, though claims by some commuters that they would return to driving into town seemed fatuous at best. NARPAC claims no particular expertise in pricing public transit fares, but it seems quite evident that Metro fares are almost certainly still too low. The chart below estimates the daily (roundtrip) cost to commute distances up to fifteen miles. Clearly, if one takes a bus for that distance, the trip can cost up to $3 dollars if an "express" line is available (bottom green line).
If one commutes by Metrorail, using a suburban parking lot for the princely sum of $3.00 a day (!) (some are $2.75), the fare ranges from a minimum of $1.35 to a maximum of $3.90 one way (at about 10 miles), and the maximum total is $13.50. Using an average in-city monthly parking rate of $13 per day, and the $0.36 per mile currently allowed by IRS for business car mileage, then the driver (unless he can share costs with another commuter) ends up paying $23.80 for a 15-mile distance (red line). Two taxi rides would be almost twice as much as that (gray line). It seems clear to NARPAC that Metrorail fares are nowhere near their limit, and that Metro parking fees could be raised significantly. As is discussed below, as Metro fares increase, the city would be wise to raise the parking rates charged in the city's commercially owned and operated parking facilities.
As NARPAC has recommended elsewhere, municipal revenues from parking fees levied on Americans' favorite possession, could be substantially higher than they are. The notion that the city can discourage car-ownership by its residents seems extremely unlikely. It is more likely to force car-owning residents to the suburbs. On the other hand, charging for "storing" cars when not in use is quite acceptable. In fact, NARPAC looks forward in the not-too distant future to seeing parking fees pegged to the "urban-friendliness" (in size, bulk, and gas consumption).
At the present time, WMATA operates (now-fully automated) parking lots
at many of its suburban Metrorail stations and charges very low rates
($2.75 to $3.00) for all-day use. The table to the right lists the stations
at which the 49,000 spaces are currently located, and compares them to
the 117,000 rush hour commuters that use those stations. The magic average
in the suburbs is 44%, and in the few DC stations with lots, it is under
20%. That these lots are currently 98.9% full on a daily basis
clearly suggests that additional parking capacity should be added. It
is also quite possible that additional parking lots to serve bus lines
could be profitable as well. The implicit assumption that bus riders don't
have cars and train riders do have cars seems dubious at best.
As NARPAC has also noted elsewhere, DC's parking rates are significantly
lower than in some other large American metro areas, on both a
daily and monthly rate. This is shown on the chart to the left. If DC
were to accept a role as a model transportation-oriented city in a model
transportation-oriented metro area, considerably higher parking rates
could be charged. And there seems to be no apparent justification for
Metro parking lot spaces to be so scarce, or so cheap, for that matter.
Metro expects to earn $21.4 million from its parking fees this year. NARPAC
thinks that number could be doubled if not tripled. People are very firmly
attached to their cars
With the valuable assistance of the Maryland Transit Administration, NARPAC has developed a chart showing the actual and planned capital investment funding obligations from FY94 through FY09. (As the budget report points out, operating budget items are actually "accrued expenses", i.e. outlays for the budget year, whereas capital investment budget items are funds earmarked to be spent on specific projects and may take several years to spend out.) Historically, these budget items have been divided between 'Rail Expansion' and 'Capital Improvement Programs (CIP)'. Starting in FY02, WMATA has divided 'CIP' into three subcategories, but they do not seem to persist into the future. The categories have been generally defined as listed below, and shown on the chart below:
o Rail Expansion which actually increases the length of track: in the '90s, most capital investment was devoted to finishing the 103-mile system, and infrastructure renewal was not yet requiring much attention. Since the turn of the century, expansion has dropped off: the monies shown for 2002 and 2003 complete the Green Line extensions, build a rail yard, and buy a few rail cars. Beyond that, nothing is yet planned;o CIP, to include:
* Infrastructure Renewal is the never ending process of refurbishing and replacing existing equipment for a fixed sized system: it involves rolling stock, passenger and maintenance facilities (i.e., stations, yards and garages), the track and other structures, and various "systems" such as the complex controls that keep the high speed trains running at very close spacing, or 'headroom'. Escalator maintenance is one major bugaboo in this category.
* System Access and Capacity refers to new items that expand the overall ability of the system to handle more riders. The new New York Avenue station is one outstanding case of a recently completed item, adding a new entrance and mezzanine at the Ballston Station is a newer and smaller project that peaks in 2005.
* System Expansion is a separate category that funds the initial engineering studies and concepts leading to system expansion such as extending the Blue Line from Addison Road to Largo Town Center. That actual project, now in the development stage, will eventually be funded by the State of Maryland and by Federal funding.
The actual spending now projected for the next five years is simply disastrous. Whatever the reason, politics, inadequate planning, lack of a dedicated funding source, or improper balance between jurisdictions, the fact remains that there is absolutely no acknowledgment of what will happen to this major transportation system unless it is continuously and promptly maintained , renewed, and expanded. The background bands on the chart above indicate the levels of spending for expansion and CIP that NARPAC believes would be appropriate asnd are developed in the next section.
The second chart combines the actual and needed capital investment with what has appeared to be adequate spending for operating costs. The latter has grown in a steady and credible fashion, and presumably should continue to do so. What is important to note in the second chart is that the total for expansion and CIP should exceed that spent for operations. That is a very different balance of spending that has ever been realized by WMATA. It is probably the root cause of the many, and increasingly serious, operational and mechanical problems currently being experienced, and seriously degrading the image of the national capital region's signature transit system.
How Much Capital Investment is Needed
NARPAC has always tried to accompany its criticisms with a positive suggestion for an alternative, even though those alternatives may not survive close and detailed scrutiny. In that spirit, the table below shows an illustrative build-up of the major items in a capital investment budget and a simple (simplistic) approach to estimating their annualized needs. It is the basis for the background areas on the capital investment charts shown above. For those interested, NARPAC's proposed 25-yr Metrorail expansion plan within DC can be reviewedhere.
To a first approximation, then, NARPAC believes that WMATA should settle for no less (in 2004 dollars) than $800M annually to keep the system expanding in consonance with the metro area, and another $700M annually to keep the current system up to snuff. The nation's capital can afford no less.
Did you find this of interest?
Please give us your FEEDBACK
This page was updated on Nov 15, 2004
© copyright 2007 NARPAC, Inc. All rights reserved