San Jose / Silicon Valley: Free Connections Make a Network

One of our recent projects was a major redesign for the bus system in Silicon Valley, more exactly Santa Clara County, in California.  The plan has been approved by the Board of Directors of the transit agency, VTA, but is stuck waiting for the BART rapid transit extension around which it was designed.

Still, the agency is moving ahead with the most critical step: getting rid of the fare penalties for getting off of one bus or train and onto another.  (This act is commonly called transferring, although I recommend calling it connecting.)  These penalties are common, but they are also insane.  Connecting from one transit vehicle to another is exactly what customers need to do in a maximally efficient network that gets the most people to the most places fastest.  Connections, in short, are what combines a pile of lines into a network.  It is insane to make customers pay extra to do the thing that uses your resource most efficiently.

The new network is a high-frequency grid system — and so, to some degree, is the existing one.  Here’s the new network, with frequent lines in black, red and orange.   (Download sharper version here.)

VTA Final Plan compressed

Wherever red lines cross, you can “turn” by changing from one transit line to the other, and because of the high frequency, the next bus or train will be along soon.  Imagine what driving would be like if there were special surcharges for turning!

To eliminate these penalties, of course, is to lose some revenue, at least initially.  So you usually have to raise the base cash fare to compensate, which sets off all kinds of alarms about “raising fares.”  VTA is raising its base fare only modestly, from $2.00 to $2.25.

But really, this shouldn’t be called raising the fare at all, because it is vastly increasing what the fare buys.  Instead of buying service only along the line you happen to be on, the new fare buys access all over the city and county.  Yes, some people who’ve built their lives around a single transit line will complain, but in such a decentralized county, with so many destinations throughout, it’s only a matter of luck if your home and destination are on the same line.  To really get places, you need connections.

22 Responses to San Jose / Silicon Valley: Free Connections Make a Network

  1. david vartanoff November 7, 2017 at 3:12 pm #

    You are spot on about transfer surcharges–too bad WMATA doesn’t understand this. IINM tthe new fare structure could also turn fares into a day pass if the required software were applied. The next step in fare decency is cumulative usage turns into a full monthly when it reaches the up front price of same. We should all remember that fares don’t significantly pay for service; they merely act as either an incentive or a disincentive to usage.

    • Dorian November 8, 2017 at 12:16 pm #

      I think my favorite example is FAX in fresno, who charges a fare of 1 dollar, transfers included (albeit paper ones). They have extremely productive service, even for a low density sprawling area, the “coverage” operating at a productivity of around 30 or 40. It is funny to me that an area that has neglected its buses, likely for decades (since the ’70s), ends up with a *much* higher performance system. Not to mention the farebox, which as I recall hovers somewhere around 25%. The VTA farebox is closer to 10%, and it’s lucky to have any “ridership” routes operating at FAX’s coverage performance…

  2. Joseph Brant November 7, 2017 at 5:26 pm #

    If fares guide rider behavior, while not affecting the bottom line much, then do so few cities have free transit?

    • Matthew L. November 7, 2017 at 6:01 pm #

      You can turn the levers of fare like VTA did and have only a moderate impact. But even small cities’ systems can make back 10% to 25% of their operating costs in fares. Eliminating fares entirely means you’re either increasing subsidy (usually property taxes) by that amount (for no new service) or slashing service by that amount.

      • Henry November 7, 2017 at 10:52 pm #

        Whether or not it’s ‘worth it’ really depends on how much fare collection is costing. When fares minus fare collection gets below a certain point it’s not cost prohibitive to cut your losses and just make it free.

      • asdf2 November 8, 2017 at 8:53 pm #

        Even if fares pay for just 20% of operating costs, and funding is fixed, are you prepared to cut 20% of service to make transit free? All the while, of course, the remaining trips are going to get more crowded, as a result of more people riding, with no money anywhere to add more trips.

        Of course, for crowding reasons, it’s inevitable that all the cuts will end up being nights and weekends, so the rush hour trips don’t leave people behind, so the end result is something along the lines of 7-day service becoming Monday-Friday only, or 6 AM-10 PM service becoming daytime-only. That’s about how bad elimination of 20% of service is.

        • Joe Brant November 20, 2017 at 11:07 am #

          For the poorest and most vulnerable citizens, saving that $5 a day could make a real difference in their lives. Ideally the county could just raise taxes on the highest earners to offset the fare revenue but we Americans seem to think of progressive taxation as the worst kind of persecution. Still, we have no problem with public schools and libraries that do not charge user fees.

      • Kenny Easwaran November 9, 2017 at 2:42 pm #

        Don’t the subsidies paid by different regions and municipalities vary by quite a lot? I would think that plenty of them are more than 20% higher than plenty of others. It seems there a way to turn a low-subsidy town with 20% farebox recovery to a high-subsidy town with free transit. I don’t think crowding is particularly an issue in most cities other than a few big ones (perhaps I’m wrong about this), so the induced demand from free fares seems like it would be a pure good, rather than a cost for the agency. Especially given that lack of fares would cut dwell times and improve driver morale (from not having to deal with potential riders that are short of fare).

    • Henry November 7, 2017 at 10:51 pm #

      For areas larger than, say, a small city or a college town, the required subsidy is too high, and on top of that removing fares removes any sort of demand inhibition for riding transit. Just like people abuse free roads, they abuse free transit, and usually it’s at the expense of walking or biking; transit is already cheaper than driving, so fares are not the primary reason why people would drive rather than take transit.

      Better to shovel the sums of money required to make transit free into subsidized transit for poor people and an expansion of service.

      • Novacek November 8, 2017 at 10:50 am #

        Also the tragedy of the commons. Damage and vandalism can go up when fares are completely free.

    • Dave November 14, 2017 at 11:47 am #

      Agree with the arguments put forth by others re: why more cities don’t eliminate fares completely, namely that (1) to do so would require cities to raise taxes dramatically and/or cut service significantly to make up the 20% or whatever gap that fares currently provide towards the transit budget and also that (2) people abuse anything that is free and free transit would simply pull people away from biking or walking which is even better for reducing pollution and road congestion than using transit is.

      Also though, I think a lot of cities charge fares simply out of inertia. The original operators of transit were profit-seeking and charged a fare, so when they went out of business and cities took over transit, they simply kept fares around out of a way to minimize the burden on the taxpayers of suddenly taking over transit operations. Path dependency kept them charging fares ever since, as nobody running the city wanted farebox recovery ratios to fall on their watch from whatever they were at the time of municipal takeover. So if that first year of municipal operation, the buses earned 20% of their cost back in fare revenue, well then, that established the baseline for all future operations to be judged by the bean counters in city hall. (This is why there is no standard for how much any particular city “should” be earning back from the farebox: some cities aim to earn back as much as 50% while others are satisfied with 10%.)

  3. Aaron November 8, 2017 at 9:37 am #

    “Yes, some people who’ve built their lives around a single transit line will complain”

    You betcha, I chose where I lived based on access to a direct rail line that went from walking distance to my house, and walking distance to my house. VTA is getting rid of that, so I suppose I’ll move again. Happy to see VTA is providing free transfers now though.

  4. Novacek November 8, 2017 at 10:48 am #

    So is it timed free transfers within a window?

    Or is the new $2.25 fare effectively a day pass?

  5. asdf2 November 8, 2017 at 8:48 pm #

    I take it the reason why the map only shows the frequent network on weekdays is because they’re too embarrassed to show the frequent network on weekends, because it’s probably much, much smaller.

    Alas is a big problem public transit faces in Silicon Valley – if you need to get anywhere on the weekend, you need to own a car. Once you already own the car, there’s no reason not to drive it to work every day when gas and transit fare cost about the same and parking is, of course, free. Also worth noting that, even with the frequent network, “turning” by switching lines still eats up 5-15 minutes of wait time. The only time turning a corner in a car takes anywhere near 5-15 minutes is during periods of total gridlock, when the buses of course, wouldn’t move either.

    Another big problem the area faces is a general lack of sidewalks and crosswalks, which makes the “foot” part of any transit journey unnecessarily difficult. I don’t live in the area, but I have had business trips in Mountain View and Sunnyvale, and I can tell you all sorts of stories about sidewalks abruptly ending and lack of crosswalks forcing long detours to cross a street. Fortunately, in many cases, it’s possible to avoid the worst of the problems by cutting through parking lots (which are technically private property, hence Google won’t acknowledge them when route planning).

    I can only hope that downtown San Jose (which I haven’t actually been too) is more walkable than this.

    • Dave November 14, 2017 at 11:51 am #

      “Also worth noting that, even with the frequent network, “turning” by switching lines still eats up 5-15 minutes of wait time. The only time turning a corner in a car takes anywhere near 5-15 minutes is during periods of total gridlock, when the buses of course, wouldn’t move either.”

      This of course is another argument for why transit transfers (or “connections” as Jarrett terms them) should be free: the rider is already paying a stiff time penalty for turning off one route and onto another so he/she shouldn’t also be asked to pay a stiff monetary penalty either. It’s not the rider’s fault that the city didn’t make the route going by his/her house also go by his/her destination. When it comes to driving, the city doesn’t penalize the people who have to use multiple roads to get from home to work because the city didn’t build a single road between the two, so why should bus riders be penalized (beyond the temporal penalty already mentioned)?

  6. Jason November 9, 2017 at 1:36 am #

    The problem is that there just isn’t enough service in silicon valley to go around to make it generally useful for most people no matter how you optimize it. The frequency of many lines are once every 30min – too low for transfers, and there are large swaths of transit deserts right in the middle of the valley. Compared to Vancouver, BC, it is downright ridiculous – even Burnaby, with similar population density, is much better.

    Furthermore, the road network is just downright ridiculous for sane network designs. For example:

    • Dave November 14, 2017 at 11:55 am #

      Gasoline in BC is more expensive. Also, free parking isn’t as abundant and urban freeways are almost nonexistent. Finally, salaries aren’t as high in metro Vancouver as they are in metro San Francisco. The combination of all that makes transit much more attractive relative to the car, even in “similar population density” Burnaby than it is in Silicon Valley… which in turn means TransLink can run buses and trains at higher frequencies and still meet its farebox recovery targets than San Jose can.

  7. hU0N November 9, 2017 at 3:14 am #

    But is it true that free connections actually reduce revenue or cost recovery?

    I don’t have access to the statistics or research to tease this out, but it seems to me that transit agencies can approach the 90° journey in basically four ways:

    • Add an additional bus route that makes that turn, giving providing a 1 seat ride on the 90° journey.

    • Allow a free connection between a bus running the first leg and another bus running second 90° leg.

    • Allow a paid connection the same as above.

    • Make no provision for the 90° journey so that riders wanting to travel this way need to drive/bike/ walk instead.

    If we count the operating cost of the whole network in units of the average fare colllected, it is obvious that the first option is easily the most costly, and costs decrease as you move down the list (number three is cheaper because people paying a transfer premium pushes up the average fare collected; it’s unclear to me where the “no service” option fits cost wise).

    But, the first option would also have the largest positive effect on system ridership, and this positive ridership seems like it too should decline as you move down the list.

    In short, it seems to me that removing fare premiums from connections should increase costs yes, but it should also increase patronage which increases revenue. So is it definitely the case that revenue goes down when transfer premiums are removed?

    • Dave November 14, 2017 at 12:04 pm #

      If the transit operator has unlimited resources, what you say is true: the first option where the vast majority has a 1 seat ride between any point A and point B should result in the highest ridership. But of course in reality, the transit operator won’t have unlimited resources, and at every point where a straight line bus route could turn, the operator will face the following choice: Send some % of the buses straight forward with the remaining buses making the turn. Each subsequent turn will cause more and more dilution of the headway (or frequency of the route), eventually resulting in a network of lots of routes going all over the place, but all at terrible headways. Almost every transit operator quickly finds that the terrible headways chip away at the overall ridership faster than the fulfillment of “1-seat rides for all” adds to it.

      • Dave November 14, 2017 at 12:06 pm #

        Promoting free transfers allows the transit operator to conserve limited resources while also maximizing coverage AND frequency. It’s the same reason why airlines gravitate to a hub-and-spoke model versus just running planes from every city they service to every other city they service.

        • asdf2 November 16, 2017 at 10:16 pm #

          In recent years, that has started to change somewhat, with renewed emphasis on new nonstop connections.

          However, air travel and bus travel are two completely different beasts. Traveling by bus, the overhead of switching lines is usually a matter of minutes – by air, it’s a matter of hours. Also, with air travel, frequency becomes much less important since people normally purchase tickets ahead of time for a specific departure. Only in cases where a flight is canceled and you have to make an emergency rebook does “frequency” really matter in the world of airlines like it does with buses.

          That said, in spite of all the above, the need to make connections is still common, and airlines are aware of it and structure their ticketing process to make it as easy as possible. Because, at the end of the day, if you want flights every day, rather than a few times a year, it is simply not economical to have nonstops between every conceivable trip pair. And people will put up with the connections for 3000-mile trips where other options, such as driving is simply not a feasible option.

  8. Jason November 15, 2017 at 3:09 am #

    Meanwhile in Sprawlville, Japan, a PRIVATE bus network looks like this:
    System map:
    Time table (red = weekday only):
    Google Maps:,143.1819903,14.5z