In Post #358, I discussed what I saw as the single most important finding of the Town’s “Joint Maple Avenue Corridor Multimodal Transportation and Land Use Study“. This post discusses the rest of that study, as presented at the Town Council (and PC and TSC) joint work session on 8/19/2019.
To cut to the chase: There’s nothing (or almost nothing) useful in the rest of the report. Not just because much of the detailed analysis was just-plain-wacky. It was, as I hope to discuss in a later post. (Example: Let’s let people park on Maple.) But mostly, the basic approach was fundamentally wrong, in a way that prevents the Town from using the results to make rational decisions about Maple Avenue. At best, I guess you might call it a place to start. Or maybe a relatively inexpensive mistake, so that you know how to try to structure a usable study.
This posting is only about big-picture overview issues. It’s already too long as is. If I want to talk about the details, I’ll have to do yet another posting.
First, I’m posting my recording of this joint work session.
Why? See Post #260. Hope I’m wrong about that, but … just in case, my recording of the 7 PM 8/19/2019 work session is at this Google Drive link.
There’s no index, because it’s only 45 minutes. Audio is lousy because there was poor microphone discipline, so it’s a mix of amplified and unamplified sound. That requires a lot of post-processing (amplification, noise removal, compression) just to make it audible. The heavy post-processing and low original volumes left a lot of artifacts in the recording.
But if you want to know what was said, and you weren’t there — sadly, looks like this is your one and only opportunity to do so. See Post #260.
The study in brief
At the 8/19/2019 7 PM joint work session of Town Council, Planning Commission, and (I think) Transportation Safety Commission, this was a Powerpoint presentation summarizing a study by Kimley-Horn. Kimley-Horn is a nationally-known engineering/consulting firm that does, among other things, traffic studies. (In fact, Kimley-Horn not only works for the Town in looking at Maple, they also work for the MAC developers, a point which caused one citizen to raise conflict-of-interest questions. Which the Town (of course) dismissed.)
Kimley-Horn looked at ways to improve pedestrian, bike, bus, and car travel on Maple and (sometimes) nearby streets. For whatever reason, they focused some parts of the study based on answers from a non-random set of about 150 people who answered their questionnaire. Otherwise, they looked at individual improvements that might be made, each in isolation. So, e.g., improvements to individual intersections, maybe moving/closing some roads, maybe adding some traffic lights, redoing some intersections, and the like. Maybe a new traffic light or two. Filling gaps where no sidewalk exists. Maybe adding some more local bus service. All kinds of things, from stuff that would cost a few dollars, to stuff that would cost on-order-of $20M.
In addition, they provided an estimate of additional rush hour trips that would occur on Maple under some future partial build-out under MAC, as discussed in Post #358.
Cost benefit analysis part 1: Wrong.
Let me start by asking a hypothetical question to illustrate a fundamental problem with the Kimley-Horn study. Let’s say you have three options for Certificates of Deposit (CDs) that you might consider investing in, on behalf of the Town. Here are your options:
- A: a CD costing $200,000
- B: a CD costing $400,000
- C: a CD costing $800,000
Think really hard now: Which CD should you choose, and why?
If you said, “But wait a minute. You have to tell me what they pay”. If you asked “What’s the interest rate”, or more generally, if you wanted to know what the benefit was that you would get in return from those costs, then you grasp the most basic problem with this study.
The consultants put together lists of things the Town could do. And they provided an estimate of costs. And that’s it.
Putting aside whether those things were at all practical (most were not), they provided no estimate of benefits or likely effectiveness of each recommendation. None. No clue as to expected reduction in traffic congestion, or in traffic accidents, or in expected pedestrian or other injuries. No estimate of the number of car trips that would be diverted to foot/bike/bus trips. No nothing. No estimate of benefits, period.
Let me give a practical example of what can happen when you do that. Suppose the consultant tells you that, for a mere $100K expenditure, you can make some intersection much safer for pedestrians. Sounds good, right? I mean, who could be against pedestrian safety? OK, suppose I then tell you that pedestrians never use that intersection, and never will. Now that doesn’t sound so good, does it? Now you realize that what sounded so good a second ago is actually a complete and total waste of money. That is an illustration of why you need to have an estimate of benefits, to complete the cost/benefit ratio, in order to select projects.
So if you think my list of CDs was more-or-less useless, then you grasp why this report was not useful, from the outset. Even if the suggestions had been plausible, the Town could not have made any rational decisions about spending money on them. Because — at least based on the Powerpoint — the Town has no estimate, not even some vague guidance, on what the likely impact would be. They got the exact analog of the list of CDs above. For any given project, they didn’t know if spending $1M would get one car off Maple, or if spending $1 would get ten cars off Maple.
Cost-benefit analysis part 2: Right
To an economist, this one is simple: What you want to see is cost/benefit analysis. At the end of the day, you want an ordered list of options for Town Council to choose from, sorted from most to least bang-for-the-buck. That is the only way that Town Council can make a set of rational decisions about what to do. And anybody who tells you differently — who tells you that you can get by without this — doesn’t really understand how these decisions should be made.
Why am I so sure this is right? The availability of money is the overriding constraint. It’s always the case that finding the money to pay for the improvements is the stopper. Of so, then for a given amount of money to spend on “multimodal” improvements, you get the most improvement by starting at the top of this sorted list, and working down it until all the money is gone.
If you do it any other way, then it’s absolutely equivalent to taking $100 bills, raking them into piles, and burning them. This is sometimes hard for people to grasp, but it boils down to this. Suppose you do it your way, pick projects that you like, and get 1000 cars off Maple, for a certain amount of spending. Then I’ll do it as I specified (i.e., efficiently), and I’ll get 1000 cars off Maple for less money. (Because? Because I did the projects in strict descending bang-for-the-buck order.) At the end of the day, to make my result exactly match yours, I’d have to take the difference in spending, convert it to $100 bills, and burn it.
Now, that’s the basic idea. But you have to face reality. Can you estimate benefits perfectly? No. Are there different dimensions of benefits (e.g., safety versus less congestion)? Yes. Do specific dedicated sources of funding sometimes mess up this neat little approach? Yes, but not as often as you might think.
Are any of those reasons NOT to do it this way? No, no, no. Why “no”, if there’s guesswork involved? Because taking an educated guess — particularly having your experts take an educated guess — is vastly better than doing things at random. And having some structured process in which to make all those guesses, under some set of methods, really beats picking projects at random. And “random” is what you get, at best, if you don’t put in place some formal process for estimating the cost/benefit ratio of each project.
I have done laundry-list studies like this myself, professionally. You MUST help the client sort the laundry list of projects, or the client simply flounders around once you hand them the report. They literally don’t know what to do first. Worse, the client will pick some order for doing projects that is totally divorced from any notion of benefit. E.g., they’ll start doing them from least expensive to most expensive. Which is not a good way to spend the taxpayer’s money.
Major problem #2: Demand is often the most important constraint.
Really, this is just an expansion of my pedestrian crosswalk example above. It’s more about one of the nuts-and-bolts of cost-benefit analysis. In a nutshell: Sure, you can build stuff. But if few people use it, the benefit is negligible. And, more to the point, you can’t really force people to use it, just because you provide it. There has to be demand for what you are offering.
Engineers are trained in engineering, not in the social sciences. So “demand” in general is not something they traditionally focus on. You ask them to look at multimodal transportation, and you’re going to get a study focused on your physical environment. And often times — as in the case of this report — there’s more-or-less no consideration of behavior. Not just of the demand for some new product (a bus line, say). But not even any consideration of how many people currently do, or might want to do, what your new infrastructure would be built to accomplish.
Here, I’m just going to give a couple of examples of where this report did not optimize.
Pedestrian safety. The pedestrian section appeared to be entirely about pedestrian safety. (And it was not about ways to get more people to take more of their Maple Avenue travel on foot, which surprised me.) Now, if you were going to ask what we should do about pedestrian safety on Maple, what’s the first thing you should do? You should ask where we have problems now. The data are readily available, as shown in Post #214.
Know what? The people who run this town already had the good sense to do that. We got the Madison HAWK light, at least in part, because that’s a hot spot for pedestrian injuries. Three people were hit by cars, in a span of just a few years, trying to cross Maple at that spot.
This continues a long-standing tradition in Vienna of looking at our own accident data to try to find places for improvements. You can see the Town’s 2010 traffic study, analysis on the Town’s website, at this link (.pdf). And you really need to tip your hat to the Transportation Safety Commission of that era, because they had none of the on-line data that are available today. They did that one the hard way, by tabulating police reports.
Barring that, if you think that (e.g.) the accident data are too sparse to be useful, then you should at least look up some estimates of what intersections get a lot of pedestrian activity, and focus on those first. Because if you’re going to put pedestrian safety improvements in somewhere … all other things equal, you might as well put them where people use them.
Either way, we are talking about assessing demand first, then suggesting improvements where there is demand for them. “Demand” in the sense that people habitually jaywalk (Madison), or demand in the sense that pedestrians at least routinely use that road crossing.
Bus service. I have made a point of riding the Fairfax Connector bus when I’m going up Maple Avenue to run errands (Post #223). It’s quite nice. It could not be more convenient (except more frequent service would be good.) And yet, I bet nobody reading this has ever ridden it, even once. Not even the elected and appointed Town officials who were talking about multimodal transport along Maple. Let alone ride it routinely.
That, by itself, should tell you something. You can supply all the nice, clean, convenient bus service you want. At great expense. But, as economists say, you can’t push a string. If nobody demands it — if nobody in Vienna wants to ride it — then that money is a dead waste. Right now, for $2, you can take a clean, air conditioned bus for a round trip up and down Maple. Runs every half-hour M-F, every hour S-S. And people ignore it.
Sure, you can recommend that the Town (e.g.) improve the bus shelters. And you can recommend that the Town run a Maple Avenue circulator bus. (Which we would not, it would almost certainly be smarter to pay Fairfax to run additional Connector service within Vienna rather than gin up our own bus line.)
But until you have some clue as to why people don’t use the bus service we have now, you might as well be throwing darts as a way to pick projects. Or raking up piles of $100s and burning them. Because you have utterly and totally failed to assess the demand for your service.
If you look at the survey results in the Kimley-Horn presentation, for what the are worth, only a tiny fraction of respondents wanted either better bus stops or more bus service. That’s completely consistent with the observed lack of Vienna ridership on the Connector bus.
So what I’m suggesting is really just common sense: Do some market research before you launch your product. Ask people why they don’t use the bus service we have now. A survey of 400 randomly-selected households ought to be adequate.
In addition, another business saw applies: Investigate what your competition does before you launch your product. Study the existing bus routes — as I have — and you will realize that they embody a lot of logic. And you will also soon realize why a trolley up and down Maple is just about 100% useless. And it ain’t rocket science — almost nobody lives within easy walking distance of Maple. If you look at how the local professionals provide bus service, it’s all about finding efficient ways to get a bus stop within walking distance of common origin and destination points. It’s not about running a bus up and down Maple and hoping there will be some demand for it.
And once you understand that, “local bus service” is a much more complex and costly proposition. Not just for the Town, providing the service, but also for the riders. If you want people out of their cars, and on the bus, you need some way to get them from their wide-spread neighborhoods to that Maple Avenue bus line.
I could go on a rant, but let me stop there. Trolley sounds so jolly. And if you’re talking about wandering around a beach resort, say, sure, that’s great. If you are talking about (e.g.) grocery shopping on Maple, and getting back home — maybe not so much.
Major problem #3: Focus solely on reducing congestion on Maple Avenue.
I thought that large sections of this report — again, as presented in the Powerpoint — were simply off-base. I’m going to explain that briefly, then call it a day.
What I thought this report was supposed to be. I thought that this report was going to identify ways to reduce traffic congestion on Maple Avenue. And, unlike other reports focusing solely on cars, this one would include using travel by bus, foot, and bicycle with a goal of reducing traffic congestion on Maple. (Although, somewhat oddly, the study does not explicitly consider the most pressing multimodal issue facing Vienna, which is rental electric scooters and bikes).
Instead, what we got was a look at the roads, sidewalks, and trails near Maple, with no single focus. Some of the report seems focused on reducing traffic congestion. But I would say that about half the analysis of road intersections was about pedestrian safety at those intersections. Laudable, but has nothing to do with reducing traffic. The bicycle section was mostly about how to get bicyclists off Maple and onto nearby streets — but nobody bikes literally on the Maple roadway anyway.
So, for example, I would have thought that the bicycle section would have been focused on ways to get people to substitute bike trips for Maple Avenue car trips, and so reduce the total number of cars on Maple. But it wasn’t. Putting aside all the other issues I had with the bike section, that was not the focus at all.
Similarly, I’d have thought that the pedestrian section was about the potential for getting people to walk their errands along Maple rather than drive. But, instead, it seemed entirely about pedestrian safety. Again, nice, but in fact, Vienna already has a good pedestrian safety record. And, honestly, I don’t think it’s fear of being run down that keeps most walkers off the Maple Avenue sidewalks.
So my final big-picture comment is that this study was unfocused. Basically, they looked at the infrastructure and suggested things that might be done. Some of those, plausibly, might reduce Maple Avenue traffic congestion (but we have no idea by how much). But many of them were just suggestions in-their-own-right, not focused on the primary goal of reducing traffic congestion on Maple.