After borrowing $1T last quarter, the US Government can now borrow for free

Take a look at this.  The yield on the 3-month Treasury bill, an instrument the US Government uses for short-term borrowing, has dropped to zero, or a couple of hundredths of a percent away from zero.

Who are the people willing to lend the US Government money for three months and want nothing in return other than a guarantee they’ll get their money back? 

The only thing I could think of is if my currency is decreasing in value fast, like Zimbabwe.  I can’t think of any reason I’d lend my money to someone for free considering it’s already in US dollars.

Minnesota Ballot Recount: How hard is determining voter intent?

I enjoyed taking a look at these challenged ballots from the Senate race in Minnesota.  For some of these, it’s really hard to try to figure out what the voter wanted.  Especially the voter who bubbled in next to Al Franken, sort of crossed the bubble out and then wrote “Lizard People” in the write-in block just like he had for some other ballot contests.  Other challenges are just asncomical, where a voter bubbled in next to both Al Franken and Norm Coleman, then drew a single line through Al Franken’s name.  This one’s a simple vote for Coleman, right?  Franken’s team challenged the ballot, stating that the voter intended to underline Franken’s name (the line goes through the center of “AL FRA”, then slopes down underneath “NKEN”).

Bubble in next to the name.  If you can’t do that, make it really, really clear what you meant.

Here’s a slideshow of the challenged ballots, but it’s more fun to vote on how you’d rule if you were an elections judge.

Source: Minnesota Public Radio, hat tip to electoral-vote.com

CVS to start selling Smartrip cards

At this week’s meeting, the Metro Board of Directors will consider a proposal to allow CVS to sell Smartrip cards at 187 area stores.  A Smartrip card preloaded with $5 in value will be available for $10, the same price as at many Metro stations with paid parking garages.  CVS has agreed to sell Smartrip at no cost to Metro and at no additional cost to customers.  CVS will display Metro-supplied register signs, and Metro will advertise on its buses and in the Metrorail system that the cards are available at CVS.

Another win-win for Metro, CVS and riders.  With the upcoming elimination of paper transfers, customers will need a convenient place to get a Smartrip card to continue transferring from bus-to-bus for free.

Crossposted on Greater Greater Washington.

Metro to share parking with new hotel

At this week's Planning, Development and Real Estate Committee meeting, the committee will consider a proposal to enter an agreement with Marriott, who will be constructing a 162-room hotel on the Prince George's Plaza Metro station site as part of a ground lease agreement (see map).  Originally, the site was to have a free-standing bank, not a very high-density use of land near a Metro station.

View Larger Map

In order to secure a loan to build the hotel, Marriott needs to show its lenders that there is adequate parking for guests.  Some of the hotel's parking needs are being filled by an agreement with the existing retail parking garage, but they also need some additional spaces for guests.  Rather than build additional parking, the hotel has proposed using 45 of Metro's 1,068 parking garage spaces, which are currently only 51% occupied during the peak period.

Marriott will pay to install exit gates that will allow hotel guests to leave using hotel cards, and will pay Metro 150% of the current day parking rate per guest that uses the lot.  Spaces will be first-come, first-served, not reserved for guests.  If the parking lot starts getting full during the peak period, Metro has the option to cancel the agreement.

The presentation also has some great maps showing the amount of proposed transit oriented development at this metro station.

Image from WMATA.

I think this is win-win-win.  Metro gets development on its property, more riders (the station is only 8 stops or 20 minutes away from downtown), and revenue from an underutilized parking garage.  Marriott reduces its cost to build new parking spaces and gets to promote its hotel as being convenient to Metro and shops, potentially increasing the room rate.  The hotel customers get parking that's convenient to both Metro and their hotel.

Crossposted on Greater Greater Washington.

Mount Pleasant 1D05 Parking Supply Estimate

As part of the debate over Washington, DC’s resident permit parking program, the issue of maximum allowable permits came up during the recent Performance Parking Hearing chaired by Councilmember Graham.  I decided to take a look at a typical neighborhood in Ward 1 to see how many street parking spaces there were compared to the number of homes.  I selected advisory neighborhood commission single member district (SMD) 1D05, because from the overhead view it looked to be almost completely townhomes, simplifying the counting.  Here’s the google map I created for this study:

View Larger Map

Results:  ANC1D05 has approximately 1.7 on-street parking spaces per household.  If every household were to receive one RPP permit and park one car, there would be only 387 left to share among the 529 homes and their visitors.  This result is likely applicable to most DC townhome neighborhoods.

Method:  Using Google Maps, I drew a boundary line around ANC1D05 using the ANC map located here.  I divided up the district into 8 sub-districts and counted homes in each sub-district, counting 529 homes total.  There were a couple of buildings that I was not sure were residential buildings, so I counted them as “homes”.  The result may be that more than one family occupies the building, meaning that the results will be non-conservative (fewer parking spaces per home than estimated).  The vast majority of buildings in the district appear to be townhome-style residences.

I then used the Google Maps scale to calibrate a ruler, with every 20 feet of curb length equal to one parking space.  I measured each sub-district’s block faces, assuming that the block face opposite the street (across the SMD boundary) would be used for parking other SMD’s cars.  For each block face, I counted the length of the block face without subtracting any length for corners or curb cuts, so that is one source of potential error in that it would overestimate the number of curb parking spaces.  Since there was not an adjacent residential area on Adams Mill Road NW, I counted both sides of the street as having available curb spaces.  The similar segment of Irving Street NW was also double-counted.  I did not count the short segment of Mount Pleasant Street since it appeared to be commercial.

My estimate of 20 feet per car is somewhat non-conservative, since a 2009 Honda Accord can be parked in 18.5 feet (assuming one foot gap on each end), and a 2009 Honda Civic can be parked in 16.6 feet with the same gaps.  These representative vehicles show that the average vehicle length and parking habits in a neighborhood can vary the available spaces by 10-15%.  This error is similar in magnitude to the error in assuming there are no curb cuts and that parking is allowed all the way to the corners.

My estimates of available parking spaces show that the sub-district between Adams Mill Road and 18th St NW, Irving and Kenyon Streets NW (“block 6”) is the most adequately supplied with street parking, at 2.6 spaces per household, while the adjacent sub-district between Irving, Hobart and Mount Pleasant Streets NW (“block 8”) is the most under-supplied, at only 1.2 parking spaces per household.

The data show that on average, there are 1.7 available curb parking spaces per house in the SMD.  Since the development pattern appears similar to other townhome neighborhoods in the District in terms of street grid density and average townhome size, the results appear to be applicable to those other neighborhoods, such as Capitol Hill, where the development style is townhomes and curb cuts are not prevalent.

The data is available upon request. 

Metro's best- and worst-performing bus lines (and how they decide)

Metro produces a bus productivity report annually. As part of the report, Metro highlights the "Worst Performing" lines, ones that don’t meet the following criteria:

1. At least 300 passengers per day
2. At least 1.3 passengers per revenue mile
3. At least 10 passengers per revenue trip
4. Less than $4.80 subsidy per passenger
5. At least 12.4% of costs paid by passengers

Hopefully once the tracking technology is fully functional, WMATA will include a 6th criterion, on-time performance.

So, which are the bus lines are identified as Metro’s “Worst Performing”?

One missed all five criteria:

24T: Heading up the pack by not meeting any of the categories is route 24T. This weekday rush-hour only service makes 11 round trips per day between the East Falls Church metro station and the Tyson’s Westpark Transit Station. It takes about 27 minutes to make the one-way journey of 8 miles. From the look of the schedule, two buses are dedicated to this route, at a subsidy cost of around $300,000 in 2007. Each of the trips averages less than 10 passengers, and fares only pay for 11.6% of service costs. With the exception of bus service along Westmoreland Street in Arlington and Fairfax counties, and Lewinsville Road in Fairfax, service along the route might be provided by other bus routes. Some of this alternate service is not really a good substitute, but for service between the transit station and the Metro, Fairfax County operates two bus lines, the 425 and the 427, from West Falls Church station. The alternate bus lines are shorter, more frequent, and operate along a limited-access highway so it’s a lot faster (9 minutes).

Missed four out of five:

98: Chairman Graham mentioned this one specifically at the parking hearing when the topic of transit to Adams Morgan came up. He’s also proposed replacing it with limited-stop circulator service, which would be faster and funded by the District instead of WMATA. This line is intended to shuttle folks from the Metro stations at U Street and Woodley Park to the hot night spots on U Street and Adams Morgan. It runs all day and late into the night (3am on Friday and Saturday), has a cheap 25 cent fare, and runs pretty frequently (every 10-13 minutes from 6am to midnight weekdays and Sunday, and all the Friday and Saturday hours too). I think this one is a transit rider’s dream in terms of service (cheap fares and frequent, regular service at all hours), but it doesn’t meet Metro’s effectiveness criteria at a subsidy cost of $730,000 per year and $6.80 per passenger. It failed all the criteria except #2 above, and got pretty close to meeting #1. It’s low fare probably is one of the reasons it has one of the lowest cost recovery of all, at 9.5%.

E6: I don’t know much about this line. It runs weekdays between Friendship Heights Metro and Rock Creek Park, serving the Knollwood Retirement Home. This one just barely met #1 above, but only had half the criteria’s passengers per trip, at 4.8. Each of those passengers costs a subsidy of almost $6.00, and fares pay for only 10.6% of costs.

Missed three out of five:

This category has too many for individual write-ups, so I’ll just list them:

• 15K/L (missed 2, 4 and 5 above)
• 20F/W/X (2, 4 and 5)
• 17A/B/F/M (2, 3, and 4)
• N8 (2, 4 and 5)
• 18E/F (1, 2, and 4)
• 3T (2, 4 and 5)
• C7/9 (2, 4 and 5)

Now, I can’t discuss WMATA’s worst performing bus lines without discussing the best. So here are the “winners” in each of the five criteria (ignoring special school service lines that WMATA isn’t allowed to operate any more):

• For passengers per day, it’s the 30/32/34/35/36 combination, at 15,500. This line had a big customer experience enhancement last year, including the addition of limited-stop “express” service, so I’m looking forward to getting new data to see how much better it’s gotten. This bus line grouping is the way WMATA reports the data. If the 70/71 and 79 were all combined, that one would be higher, at 16,830. The X1 and X3 are sister lines to the X2 but they don’t bring up the X2 enough to take the top spot here (total of 14,071). The 90s lines don’t really combine (96/97 don’t follow the same route as 90/92/93). For the rest of the categories, I’m just going to report the highest based on WMATA’s data since it would be hard for me to combine cost recoveries or passengers per revenue mile.
• For passengers per revenue mile, it’s the X2 at 10.3. This radial line connects the Minnesota Avenue metro station with downtown along one of DC’s busiest main streets. The line is on the short list for being upgraded to streetcar service.
• For passengers per trip, it’s the C2/4 at 66. This cross-suburb line connects Twinbrook, Wheaton, the University of Maryland, and the Greenbelt metro/MARC station. It’s a long line and it connects a lot of destinations to Metrorail, so it’s not a surprise this one has a lot of passengers per trip. It ranks among the top lines in other criteria, too.
• For subsidy per passenger, it’s the X2 at 56 cents. Interestingly, even the best bus lines don’t operate at a profit.
• For cost recovery, it’s the 11Y (an express route direct from Mt. Vernon to Farragut Square) at 56.3%. This line makes 6 morning northbound trips, and 6 evening southbound trips for a fare of $3.00. It’s followed closely in the category by the X2 again at 55.7%.

From this performance, it looks like replacing the X2 line with a streetcar is a really good idea. We’ve obviously found a route that a lot of people can use. Streetcars travel faster (due to higher torque electric motors), have lower operating costs for high ridership lines, and can carry more passengers per vehicle, reducing the number of vehicles and operators needed to serve customers. They just need a route that has high ridership to justify the fixed investment. I think that investment is justified.

The data is available by request until I can find someone to host it. Send an email to my profile address.

Transit Wayfinding

One of the interesting comments at the parking hearing I attended was from one of my panel-mates, a businessman from Adams Morgan. He stated that if people only knew that the 42 bus went to Adams Morgan from Dupont Circle, more people would ride there. Right now the headsigns from Dupont Circle only say "42 - Mt. Pleasant", because that's the ultimate destination, but Adams Morgan is on the way.

This brings up an interesting debate about what the balance should be between overwhelming people with too much information (i.e., the metro station "U Street/African American Civil War Memorial/Cardozo) and not providing enough information (i.e., the station named "Smithsonian", which is not the closest stop for many of the museums).

Another example of this issue was when I saw a family having trouble with a metro farecard machine. Turns out they were just trying to figure out which station the Verizon Center was at. Normally I would say that we should try to keep our station names short to avoid cluttering the map. When the destination is as popular as the Verizon Center and visited by many people that don't normally take Metro, we should probably add the name to the system.

Metrobus operates the 38B, dubbed the “Orange Line with a view”.  The buses feature prominent Metrorail-style graphics showing station names similar to those on the Orange Line.  This might be an example of a best practice for buses:  Show the major neighborhoods but not every stop.  Brand the bus with a specific purpose, and give people an idea of what’s going to be on the route.

What should the balance be? Should we list all the major neighborhoods traversed by a bus line, or just say "Silver Spring via 16th Street", as in the case of the S lines?