Hire In-House, Don’t Use Consultants
I recently found two presentations, one from 2017, the other from earlier this week, both underscoring the importance of in-house expertise for efficient construction. This is layered on top of interviews Eric and I did for our Boston case study and a few additional interviews I did in other American and European cities. It is my professional opinion that agencies that engage in major capital projects, even if they involve rolling stock acquisition rather than the construction of new lines, ought to hire in-house, and make sure to have long-term capital programs.
Both presentations concern rolling stock. The one from 2017 is by Stadler, regarding the challenges of the American market. On slide 32, it mentions that Caltrain was a demanding customer, with all expertise outsourced and yet managers engaging in micromanagement. The micromanagement is in line with what we’ve heard from contractors for other capital expansions, like Second Avenue Subway, especially contractors with experience in both the US, where this practice is common, and Europe, where it isn’t.
Thanks to the factors mentioned by Stadler as well as the Buy America requirement to set up a new factory with a new supply chain for a midsize order, the cost is $551 million/96 cars, or $5.74 million/car; the typical cost of a KISS is 300 million €/90 cars, and the €:$ ratio is not 1.72, far from it.
The other presentation, from this week, concerns the MBTA’s slow approach to electrifying its commuter rail network. It wishes to begin with a pilot on the already-electrified Providence Line, but is running against the problem of having no in-house expertise, just as Caltrain does not. The presentation on this says, on slide 3, that it takes 6-9 months to onboard consultants, and another 6-9 to develop performance requirements for a kind of vehicle that is completely standard in high-performance regional rail networks in Europe.
Instead of hiring experienced professionals (who must come from Europe or East Asia and not the US), the MBTA plans to piggyback on either the overpriced Caltrain order, or an obsolete-technology order by New Jersey Transit. The Caltrain order, moreover, is stretched for the generous loading gauge of the Western US, but does not fit the catenary height on the East Coast, even though European KISSes would easily and are around 13 cm lower than existing MBTA rolling stock.
Prior Northeastern examples
This combination of political and managerial micromanagement with outsourcing of technical expertise to consultants is common enough in the United States. In the Boston report on the Green Line Extension, we were told by multiple sources that the MBTA only has 5-6 engineers doing design review. Thus, they have the capacity to handle small projects but not large ones.
Small-scale projects like building a new infill station or taking an existing low-platform commuter rail station and converting it to an accessible high-platform one usually have limited cost premium: in Berlin, infill stations are 10 million € outside the Ring, whereas in Boston, infill stations and high-platforming projects (which are very similar in scope) are around $20-25 million – and Boston platforms are longer. This is also the case in Philadelphia, where headline costs are lower because the stations are smaller, but overall the unit costs are comparable to those of Boston.
But large projects are beyond the ability of a 6-person team. The required permanent staffing level is likely in the teens for a team whose job is just to score design and construction contracts. This choked the original Green Line Extension, leading to bottlenecks in design and contributing to the project’s extreme cost. The restarted version is still extremely expensive – it’s getting some good press this week for running slightly under a $2.3 billion/6.3 km budget, but said budget, $360 million/km, is well above the international norm for a subway, let alone trenched light rail. The current project has sunk costs from the previous ones, and a combination of in-house and consultant design about whose efficacy we’ve heard conflicting reports, but the team is much larger now.
In areas that don’t even have the skeletal design review staff of Boston, costs are high even for small projects. Connecticut deserves especial demerit: its department of transportation relies exclusively on consultants for rail design (perhaps also road design but I do not know), and infill stations cost not $20-25 million but $50+ million. The Hartford Line, compromised from the start, even displays this state-by-state difference: the one Massachusetts project, a single high platform in Springfield, cost $10 million/100 meters, a fraction of comparable projects in Connecticut. Larger Connecticut stations, such as those for Metro-North, have seen extreme scope creep, amounting to a $106 million total cost.
Consultants and design
American agencies speak of design-bid-build contracts, in which design and construction are separate, and design-build ones, where they are combined into a single contract. Design-bid-build is superior. But really, contracts in low-cost countries are often neither of those, but just build contracts, with design done mostly in-house. A procurement official in Stockholm explained to me that Swedish contracts tend to be build contracts; design-bid-build can sometimes be used with supplemental consultants helping with design, but it’s not the norm. Moreover, in Oslo, the use of design consultants instead of in-house design has not been good: consultants tend to engage in defensive design because of how Norway structures risk allocation, leading to overbuilding.
In Spain and (I believe) Italy, contracts are design-bid-build. But there’s so much in-house involvement in design that it’s more accurate to call these build contracts. The in-house design teams are not huge but they’re enough to work with private design firms and score proposals for technical merit. In Istanbul, the system is somewhat different: preliminary design at the 60% level is contracted out separately from the combination of final design and construction, which may possibly be called des-bid-ign-build, but the design part is extensively scored on technical merit, at 60-80% of the total weight. The construction contracts in Istanbul are lowest-bid, but contractors can be disqualified, and since Turkey has so much infrastructure construction, contractors know that they need to behave well to get future work.
Unfortunately, American consultants believe the opposite: they believe in the superiority of design-build and are not even aware of pure build, only design-bid-build. Sources from that world that I generally think highly of have told me that directly. But that is because the sort of projects that they are most likely to be involved in are ones that use consultants, which definitionally are not build contracts. The ongoing expansion projects in Stockholm, Madrid, Barcelona, Milan, Rome, and Berlin have no use for international consultants, so international consultants are not familiar with them, and end up knowing only about high-cost examples like London or the occasional medium-cost one like Paris. In effect, to rely on consultants is to ascertain one largely learns worst industry practices, not best ones.
The alternative to paying consultants is to obtain public-sector expertise. Agencies are obligated to hire sufficient-size teams, and pay them competitively. Engineers in Italy and Spain have a lot of social prestige, much as in France and Germany; even in medium- rather than low-cost countries in Europe, like France, we were told by UITP planners that the people planning metro systems are hired from the engineering elite (in France, this would be Grandes Ecoles graduates), and paid appropriately.
In the US, there is no such prestige. Humanities professors speak of STEM privilege routinely, but by Continental and East Asian standards, the US and UK have no STEM privilege: the elites are generalist and are not expected to know the specific industrial fields they oversee. The public sector thus treats the planner and the engineer as a servant to the political appointee. Senior management routinely ignores the advice of younger planners who are more familiar with present-day research.
The pay, too, is deficient. In absolute numbers, planners at American transit agencies get paid better than their European counterparts – but American white-collar wages are generally higher than European ones. The MBTA pays project managers $106,000 a year as of a few years ago, which is a nice wage, but the Boston private sector pays $140,000 in transportation and more in other fields. The public sector, through budget-cutting officials, sends a clear price signal: we do not want you to work for us.
There is another way, but it requires letting go of the idea that private consultants are better than long-term in-house experts. It is obligatory to hire in-house at competitive wages to grow the design review teams, and listen to them when they say something is desirable, difficult, or impossible. Instead of onboarding consultants, agencies should immediately staff up in-house with plans for long-term investments. Moreover, senior management should back the planners and engineers when they engage in value engineering, even if it annoys politicians and local activists. The role of elected politicians is to review those in-house plans and decide whether there is room in the budget for the megaprojects they recommend, and not to micromanage. This way, and only this way, can the United States shrink its procurement costs to typical Continental European levels.
Thank you Alon for another insightful article. Within the civil engineering world generally (CalTrans freeway projects and CA Dept of Water Resources Water Diversion and Flood Control), two issues come to mind that weren’t discussed, but also maybe drive costs up higher in USA than Asia, but not sure about this relative to Europe: 1) Design—Environmental Impact Report (EIR)—Redesign—Agency Approval—Public Review. These are required by law, so engineers and consultants try to time and control information flow so as to skirt around and past Environmental NGOs and the general public who must complain in timely manner. Public review is within the neighborhood jurisdiction where project is located, so affected downstream/upstream public in adjacent neighborhood may not be notified in time, and some public review complaints are screened out of the record because they may require substantive design improvement. However, in USA, the military has terrific power to bypass many such rules, even EIR. The US Army Corps, for example, is famous for making environmentally destructive flood control projects that in the end also failed to protect the region from disaster, such as Lower Ninth Ward of New Orleans from flooding caused by Hurricane Katrina.
There are environmental laws everywhere, but the use of lawsuits to enforce them in the US is absolutely a contributor to high costs. In Europe, too, the difference between countries that enforce environmental laws through citizen voice lawsuits (i.e. Germany) and ones that have state bureaucracies of scientists doing the same (Italy, Sweden) is notable. Germany is very NIMBY and Herbert in comments here has long complained about consequent underinvestment, like the NIMBY-derailed Y-trasse connecting Hamburg, Hanover, and Bremen.
Some of the NIMBY and Environmental problems could be avoided if feasibility and design teams were more sincere and multi-disciplinary in makeup. Right from the start Engineers often make blunders related to misunderstanding of environmental and socio-economic issues, so rather than try to comply to full intention of the law, they make a token half hearted approach based on rank ignorance, that complies to minimal extent of law, then try to hide design tech data. So, when then the law suits are filed, this drag down project, contract price rise, or maybe protect never materializes. They get paid for doing nothing, and if project gets approached against public review, results are mediocre long term.
From my experience with electric utilities, Municipal owned utilities(MOUs) tend to have an internal to contracted engineering design ratio of 80:20 while most of the Investor owned utilities(IOUs) have about a 30:70 to 40:60 ratio. This reflects in project costs with IOUs having consistently higher costs relative to the municipal owned counterparts.
The financial incentive structure within most IOUs rewards maintaining a low in-house engineering head count on the books at the expense of higher overall cost structure. Hence majority of the design for capital projects, transmission lines, substations, power plants are contracted out. The cost premiums between contracted EPC ( Engineering, Procurement, & Construction) and Internal engineering plus contracted construction projects is between 15% – 30%, but not only that, EPCs tend to require more corrective post construction work incurring additional capital expense thats not included in the ex-post project cost.
Some people within management are aware of this cost premium, but the organizational inertia means they just default to not rocking the boat and continuing with business as usual. For the finance department, they’ve done a good job if they can keep recurring head count costs low regardless of if they incur higher overall costs down the line. I don’t know if the unwillingness to carry in-house design engineers on the books is also a factor with agencies in charge of public works projects.
Caltrain KISS is 13 cm too tall? Thought I saw photos of double stack containers on NEC that makes me think clearances are higher.
No. There’s only a few places on the NEC where the wire height can take tall freight, and they mostly cluster on very short stretches where connecting freight branches from further inland are making their last mad dash to the coastal ports and end up either crossing the NEC or co-running on it for very short distances. Default out-in-the-open wire height can be pretty generous, adequate for 17 ft. “Plate F” cars (by far the most common type of high car). But the squeeze-down around old overhead bridges is where it really restricts.
Safe wire clearance over an unshielded freight car roof is +2.5 ft. for 25 kV electrification (New Haven-north & all new-construction), +1.25 ft. for 12.5 kV electrification (Washington-New Haven, Keystone Corridor, & SEPTA). With domestic double-stacks being 20’2″, that means you need 22’8″ as a catch-all for running anything under the sun beneath 25 kV overhead.
Tightest squeeze under wires in MBTA territory is 2 bridges between Back Bay Station and South Station in the combo cut containing the NEC and I-90. No freight–not even materials deliveries to the T or Amtrak–roams in that close to South Station so it’s never been an ops issue to-date. Tallest current MBTA cars are 15’6″, their slide presentation claims 15’11” is the theoretical max under-wire, and the Caltrain KISSes are 16’1″. Not sure where the 10 cm figure came from unless their slides typoed the KISS heights too low, because the discrepancy looks more like 5+ cm. The Bombardier MLV’s that NJT is ordering are fine because at 14’6″ they’re designed for the more-restrictive Penn Station/East River/North River complex, Grand Central/Park Ave. complex, and Philly Center City complex (the tightest passenger vertical clearances on the continent excepting yet-to-be-opened LIRR East Side Access). The Eurozone KISSes would be fine, but Stadler only bid the Caltrain Frankenmods for the T’s RFI so that’s all they currently have to choose from. I don’t know if that’s the fault of the RFI bid specs trapping Stadler from bidding the straight Euro import, if Stadler internally decided they couldn’t stick their necks out for another buyer unless they direct-recouped some value from their thus far tortured and expensive Cali mods, all of the above, or none of the above.
If Providence/Stoughton were the only line ever to be electrified, undercutting the trackbed at those two 15’11” spots would be an easy fix. The problem is more that the un-electrified Worcester Line has even tighter clearances in that same downtown I-90 combo cut after it diverges from the NEC, and that’s one of the highest-priority brand new wire-ups on the system to come next. The tightest bridge there is a shade shorter, but also abuts a rail overpass of a small stream so the trackbed-shaving games are no-go for attempting to split the difference. Will have to be done up as short dead section of wire to coast through as an ops-kosher kludge. After Worcester, the Grand Junction Branch linking the T southside to the northside via Cambridge has an even tighter restriction abutting its Charles River crossing which can’t take anything taller than the T’s current 15’6″ bi-levels in its present un-wired state. That one is super-iffy for electrification because neither trackbed nor bridge have any give; it can *potentially* eke by with a dead wire section, but would be a long enough dead section to potentially “gap out” a train that had to pause on the bridge for any purely chance reason. So universal deployment across the system gets a little bit harder adopting the FrankenKISSES vs. something natively a bit shorter.
Things would be way better if Caltrain’s refuse weren’t the only bid option and the un-fuckered Euro imports were fair game, but unless Stadler changes its mind in the Request For Proposals round that’s all they’re offering to market as of this moment.
Undercutting is not easy. Just buy trains that fit within the clearance envelope; many exist and the T can spec that out if it cares, which it doesn’t.
Is it non-trivial to modify euro-import Stadlers for 48″ boarding height?
RE: Brendan Dawe:
Modding a generic Euro-import KISS isn’t physically difficult. Indeed the Euros may already be height-compliant, as the T’s Board presentation seemed to indicate that it was specifically the Caltrain-modded KISSes that were over-height. The problem is that Stadler only bid the Caltrain Frankenmods for the T’s EMU Request For Information; straight Euro imports (with any door interface mods for the loading gauge) weren’t offered as any option to begin with. The RFI established the base vehicle specs to gauge level of interest from the global rolling stock market before the subsequent Request For Proposal round of bidding where the manufacturers play for keeps on the order bidding for the set unit quantities, delivery schedule, warranty/service package, and Buy America considerations for the assembly. The RFP round may now also be structured to get these ‘early’-plug cars for Providence/Stoughton, then be spread out the back-end options for the fresh line wire-ups that will take longer to enact.
If Stadler self-chose not to bid a Euro import for the RFI specs-fest, it’s quite likely they aren’t interested in bidding that for the real-deal RFP. And thus it doesn’t really matter what solution they *could* potentially offer to suit, because they’ve made up their minds about the only one they *will* offer. And it ain’t the Euro imports. That’s the crux of the problem. Coulda-shouldas have now been culled by what the vendor says they will/won’t do. There’s a couple possible reasons for this.
— The T’s RFI specs were fatally flawed, and sprung some self-laid booby trap that practically excluded the EuroKISSes from being able to be straight-imported (excepting whatever door-to-platform interface mods routinely get done to those same EuroKISSes when they are adapted for different countries). There was plenty of the wholly expected U.S.-centric bloat available in the RFI specs for critique, and indeed they did get generalist criticism for not self-policing enough on car weight. But there was no hyper-micromanagement of specs like Caltrain disastrously engaged in, so the resulting bids were not exactly ripe for “unicorn”-skewed pigeonholing. Advocacy groups like TransitMatters have had complete access to the publicly-available specs since Day 1, and other than the same generalist weight concerns and TM’s specific preference for single-level cars (…but 3 out of 6 bids received did end up being single-level) there wasn’t much specifically red-flagged in the specs. Vertical clearance certainly didn’t project as a market-limiting problem spec, because Greater Boston’s clearances are almost universally more generous than NY/NJ and Philly where the North American EMU market is most hotly vendor-contested. It’s not very clear today *what* exactly could’ve/did thin the herd to specifically exclude the import KISSes.
— Stadler internally drew a line in the sand around its Caltrain Frankenmods. While that order is now proceeding after the last of the embarassing door-to-platform incompatibilities got fixed, the manufacturer is absorbing a ton of sunk cost for how tortured that procurement was for design mods. They may have internally placed emphasis on bailing out that Cali folly with more same-spec unit orders to increase their profit margins. They may have decided that the Caltrain mods were already so intensive it would’ve driven up costs at their extant Buy America factory to assemble a separate Euro-import order there. They may have been traumatized by the amount of micro-management from Caltrain and not seen adequate enough protections from the T engaging in the same behavior, and decided the in-theory straightforward EuroKISS door-to-platform adjustments were too tempting to over-fucker with…so they chose the straight Caltrain-mod vehicle instead of even offering the option of a Euro import. It could’ve been something completely different and/or more obscure that cooled their interest. At any rate, the RFI bids were last year before the T was considering early Providence/Stoughton adoption of anything, and Stadler still only submitted a Caltrain-or-bust bid package. For whatever reason, they corporately decided that was all they were ever willing to bid here.
— Some combination of the external/specs and Stadler-internal…potentially a non-obvious reason that doesn’t stand out from eyeballing the RFI specs and then eyeballing what’s different with the Caltrain Frankenmods. If the RFI specs truly are at fault, it probably needs a deeper-dive analysis from the advocates and watchdogs to pinpoint exactly what’s objectionable and so self-limiting so it can be corrected before the RFP round. TransitMatters has now released a couple of Regional Rail corridor implementation plans that seem to rely on vehicle acceleration profiles some degree above-and-beyond what the 6 RFI bid packages will support (example: Boston-Hyannis, 79 miles in 1:19 requires acceleration braun that’s hard to see any of the bid packages being able to natively support without major spec changes). It would be nice if they put out some policy docs deep-diving the RFI specs to match and juxtapose what does/doesn’t work for those service levels, because right now there appears to be a large unattributed difference begging to be filled with specifics (not to mention a gaping question of “How we gonna get the vehicles to run this when these 6 bid options clearly ain’t it?!?”). From the market side of things, probably needs to be an inquiry made to Stadler to expand at greater length on what the differences are between an import EuroKISS and the Caltrain Franken-mods to their manufacturing/assembly capabilities, expand at greater length on what they saw in the RFI specs that business-precluded a EuroKISS bid, expand at greater length at how they see themselves breaking the North American EMU market, and expand at greater length about how Buy America insanity distorts their business model. And so on.
It’s one thing if there’s a specs obstruction inhibiting better bids. Find the offending spec and fix it. There probably is something awry in there, but it hasn’t been spelled out for us to total exactitude yet. It’s another thing entirely if the manufacturer, for wholly internal reasons, just ain’t that into you and isn’t ever likely to bid exactly what you want. Saying “Fuck it…we’re just gonna self-override and buy the imports” sounds simple enough on-paper, but if federal funding streams are going to be made available for this order it’s going to need to go through the Buy America gauntlet and all the challenges that entails (common-sense anticipation here, since the $1T fed infrastructure effort could well slush some money these vehicles’ way somewhere along the line). In the real world, those 6 bid packages they received for the RFI overwhelmingly contour the market interest in serving their actual order come RFP time. So it’s critical to find out where the inflection points are in the base specs, and also the inflection points in the likely bidders’ internal current affairs. And do it soon, because the RFP round of bids is coming soon.
Thanks F-Line. Detailed and itneresting as usual
Look, Stadler (the real Stadler, the people who do rail enginering, not the 100% overhead “Buy American” Potemkin grifters who do rent-seeking and fiscal “engineering”) pretty much exists because their design and production personnel and facilities were are remain very capable of doing crazy one-offs and customizations for pretty much any reasonable customer.
Narrow gauge? Broad gauge? Almost-unique gauge? Bring it on. Boarding height here, there, or over there? Yeah, got that covered. 25kv? 600v? Diesel? 16.7kv? Piece of cake. Whacky cog-wheel adhesion? Sure, whatever. ETCS? We’ve got that in-house, no problem.
The concept that Stadler AG the engineering concern couldn’t crank out vehicles that could provide they-wouldn’t-know-what-hit-them quality service on MBTA lines at something not a lot more than half the “Buy American” Potemkin grifter “costs” isn’t borne out by their actual global track record. Bodging aluminium body shell dimensions a bit this way or that really is something they seem to do in their sleep.
Of course nobody who knows is going to say anything, but I’m personally imagining this is a case of the client firing the client. Some business just isn’t worth having, and who the hell wants to work for insane assholes?
Thanks for that judged-by-me-reliable info on clearances, F-Line – and the Wikipedia article on Nubian Station is quite a revelation too. (I’m not familiar with Boston.) With at least three volt-amp combos on wires in the NYC region alone (I have it recorded where I’m about to place the text from your comment.) knowing what clearance is required on a given route can be so time consuming as to become futile. It may be that all of it will someday be unified – in a VERY optimistic scenario. (Surprising to hear ESA has such tight clearances – to match the steep grades.) – But I prefer the idea of switching to linear propulsion ASAP, notwithstanding having no commercially produced version of it that can go faster the 60mph at the moment.
Because of its occult (secretive) nature, contracting for rail projects avoids the light of day concerning those in charge – and scrutiny by the general public – who given only a general understanding might yet be useful in stopping some of the more ridiculous rail planning ventures expected to get etched permanently into stone and steel pretty soon around here. This despite NEPA and all the requirements for openness – clearly they’ve found ways to get around it – and this COVID of Transit Planning virus has clearly spread across the ocean to England, EVEN England, as well.
On looking at imagery from 2019 of West Haven, the station is already pretty highly developed, with 1100′ (335.28m) side platforms varying in widths around 15′ – with the inbound (to NYC) side generally wider. There appears to be a commodious brick-and-glass stationhouse on the inbound side built in the last 20 or 25 years, with an enclosed bridge to get to the other platform, and no doubt fulfillment of the ADA requirements for it. So it’s hard to determine from the Chronicle article what all this expensive work is about. It might or might not be aptly comparable to the raised platform project in MA. So here you see another example of public planning where again, the necessity of obtaining information needed to form an opinion could potentially take a lot of time.
In the case of East Side Access and its 25-year-old parent project ARC the public has been deliberately kept in the dark, and the same can be said for the political entities in charge of decisionmaking of a highly technical nature. Drawings for instance of the Amtrak Bypass Tunnel in Sunnnyside – first to be bored – then cut-and-cover – then bored again (It was bored, I happened to see the TBM being removed.) are still shrouded in mystery even as to general alignment, and the only publicly available drawing of it is really an insult, as if Parsons (or WISP or whatever) was flipping to bird at both the public and the political entities in charge. (I think this kind of “elitism” in US railroad engineering is an established fact.) Necessity of the bypass tunnel is questionable given a well intentioned and thought-out overall project, since no expense was spared in 1910 to have the two center tubes of the East River Tunnel crossing underground, east of the East River in Queens.
While it could be surmised from the information available that the project would debilitate Harold Interlocking gravely in terms of speed and train capacity – making the two-track connection onto the 4-track Hell Gate Bridge there quite unavoidably permanent – it was a big surprise this winter to spy from the train window the apparently-3%-grade of the four ascending tracks from Grand Central, which enter just west of Harold Interlocking. The explanation (certainly no justification) is something like this: There was insufficient linear space to get a decent grade located between the two cross streets which the existing 8-track line bridges over, and in order to avoid the “expensive and time consuming” job of service outages on a succession of individual tracks to connect the new bored tubes under Sunnyside Yards to something nearer the surface at the point where they enter daylight, they simply bored them as far as possible. This was remediated BY CHANGING THE ORIGINAL GRADE CONFIGURATION, increasing the grade going down towards the East River Tunnel, in order to have the surface lower where the ESA Vomitoriums from Grand Central enter the line. All that though, failed to avoid what by all appearances is a 3% grade.
It is a nightmare scenario of debilitating speed and capacity restrictions imposed by the project. And add to that the additional trains and space-consuming stopping movements of local service on the Penn Station Access Project, expected to open shortly after ESA. Were the demands on capacity of the combined projects EVEN CONSIDERED by the said political entities in charge? This is what we have done to our premier ‘high-speed’ line, the Northeast Corridor – far from improving it. The one remediating force in all this is thought to be Janno Lieber, who was able to get the notorious change orders and force accounts of ESA more or less under control, albeit after most of the damage in that decade-long chiseling racket was already done. He doesn’t come from a railroad or engineering background though (otherwise he’d be completely useless of course) and suffice it to say, they’ve got him where they want him.
Double stack trains from Los Angeles port cargo and delivery of empty containers are stacked/I stacked in Fontana, I believe, which is on eastern edge of LA metro area. Fontana is also home to nations busiest Truckstops.
As for secretive planning, law requires notification of local authorities, and in case of Benicia, Ca, for example, city was able to STOP announced plan to receive daily at troubled Valero refinery some 200 sulphur rich shale oil tanker cars from Canada, on tracks that span the Sierra Mountains where fuel spills have caused environmental disaster in the past, and delivered to refinery with history of problems https://www.google.com/amp/s/www.sfchronicle.com/bayarea/amp/Benicia-issues-warning-about-Valero-refinery-13712482.php
ESA’s clearance issues are at least forgivable because they got locked in with the 63 St tunnel, and everyone from that project is probably dead or retired at this point.
But how common are third-rail variants of modular rolling stock, anyways? I was under the impression that most systems in Europe are wired.
RFI bid specs trapping Stadler from bidding the straight Euro import,
Which one? Almost all of them the gap would be too wide. The floor heights would be a problem too.
Everyone east of the Rockies has decided that “Amtrak” loading gauge means they will be able to get something from multiple vendors forever. Settle for something the MTA is using or something NJTransit is using, their fleets are big enough that there will be support and parts until the trains wear out.
They can stretch the horizontal loading gauge easily as all vendors do for the Nordic market, it’s not that big a deal.
And actually no, what NJT is using is obsolete tech at a large cost premium, around 70%.
Going with Stadler is going so well for Caltrain, isn’t it?
A very quick glance at Wikipedia says the Swedish trains would have a gap that is too big. The floors are too low.
Cost per meter, lineal or cubic? Or per passenger?
Stadler’s website has a PDF for the Caltrain KISS EMU where the height is listed as 15′ 10.5″, so MBTA’s 15′ 11″ clearance could be met.
It’s a mystery to me where the extra two inches came from.
In the Netherlands, regional transit authorities hire consultants because bus concessions run for 10 years and the tender process takes only about 2-3 years. Even if you have a public operator, there are things like large-scale network redesigns that only happen once every few years and require a lot of expertise then, but not in between. Like Jarrett Walker does.
I guess you could coordinate the timing of redesigns and have in-house expertise at a higher level of government, but that would still basically be consultants.
For infrastructure projects there are some large engineering companies that do design work, that have “framework contracts” with the national and local governments. There are a lot of one-off projects so you can’t really hire in-house expertise. If you want to hire people with repeated yearly contracts instead of fixed contracts you have to hire people as self-employed consultants, even if they exclusively work for you for 5 years.
I bet the lack of a ‘rolling programme’ for transit expansion contributes to high costs anyway, and this consultant structure may be part of it. I’m curious how for instance Sweden deals with this.
Should the in-house team always be at a local level or would efficiencies be gained by having a federal body taking care of this, at least for cities below the level of NY, LA, Chicago? There are plenty of cities that aren’t big enough to have a viable ongoing infrastructure construction schedule that would necessitate permanent staff, so in this case they could make use of the federal body delegating out resources. Of course the Federal government right now is even more bloated and dysfunctional than state/local counterparts, but would this work in theory?
Is there any evidence that the federal goverment is actually more bloated and disfunctional than its state/local counterparts, especially as it pertains to transit?
Like being more bloated than the MTA, Caltrans or CAHSRA would be geniunely impressive.
Given the low level of activity in general around the country, a national team should be the answer. But federalism will prevent it.
I’m pretty sure you’re slightly misrepresenting your Swedish source, because major Swedish infrastructure projects typically employ design consultants extensively. Not even Stockholm expands the subway regularly enough to keep the requisite technical design staff onboard on a permanent basis.
I think when you say “keep design in-house” you really mean “keep project ownership in-house” i.e. the responsibility for project scope and risk must remain on civil servants, who must have sufficient capacity to call the shots. This doesn’t eliminate consultants completely or even to a great extent, but allows them to remain in a consultative role.
I wouldn’t say there should be no consultants. You cannot know everything. However you need enough in-house that you understand and can control the consultants. Unless you have a consistent budget over the next 30 years doing everything in house doesn’t make sense – the years when there isn’t anything to build you still need to pay your internal people – and worse you need to figure out how to keep their skills sharp. Consultants should be going to other clients and thus keeping their skills sharp. However consultants won’t know the details of your situation and so will need you to catch them up.
In order to manage consultants, you probably need to know everything consultants are talking about, or consultants would prioritize their objective (maximizing the profit) rather than represent and stand firm on clients’ interests.
Haven’t you seen people or organizations hiring a consultant just to manage consultants because they have no clue on the subject or project? I have seen a lot of this in public sector in North America.
The same applies to in house as well though. You have to manage people or they will do things that benefit them not your goals. You can hire good consultants and black-list the bad ones. You can hire good people, or just whoever. Management is not easy.
News Flash – the Hudson Tunnel Project has a Record of Decision issued today. Re: Consultants and In-House planning – the project was employing both – at NJ Transit, the MTA and the Port Authority – for most of the time since the ARC Tunnel folded in 21010. I wonder what they were doing!
The first three segments of Edmonton’s LRT (including the downtown subway) were built on-time and on-budget, despite being built in a boomtown environment. The fourth segment was late, over budget and had the only construction fatality. The first three were managed by veteran city engineering staff, the fourth was run by consultants. As mentioned in other posts, consultants can provide specialized expertise but shouldn’t be put in charge.
Interesting tidbit about consultants from Madison (Wisconsin): At a recent Transportation Commission meeting, we were asked to approve a consulting contract for our BRT project. Part of the justification for using a consultant was that the current BRT project would be over at some point. I raised that it’d still make sense to build up in-house expertise. The transit GM conceded that point but also pointed out that there’s a sizable risk for a local agency that you train your in-house experts, only for them to be poached by other cities or consultancy companies — as evidenced by the fact that he was poached by our city from Indianapolis.
Did the GM ever stop to think that maybe such poaching is a good thing because this forms multilateral ties across the industry?
It might be on an industry level, but I would have assumed you, as a mathematician, would have immediately seen that the free-rider problem makes this a prisoner’s dilemma.
It isn’t, though. The agency that hired in-house gains from the knowledge spillover too.
It would have to be a pretty big spillover to offset the cost of training someone.
You’re onto something here, but unfortunately you’re undermining your own research by overextending the conclusions that can be drawn from it.
There is a difference between design capacity and ownership capacity (i.e. the skills needed to own, operate, manage and develop) as system. It doesn’t really matter whether we’re talking about a subway, IT systems or industrial equipment, it rarely makes sense to roll your own.
Designing earthworks and bridge structures isn’t going to be a core competence of a transport agency. It’s core competence should be to transport people cost-effectively. Those same skills do help you avoid gross overengineering because they give you the vocabulary to effectively communicate your needs to consultants.
Ownership skills are also a partial prophylactic against the other major contributor to cost overruns, delays and change orders, because a major source of these is that the customer changes their requirements midway through the process. However, it doesn’t protect that much against the other major reason, that construction begins before the designs are complete. That’s more a question of proper risk analysis, allocation and communication.
Patrick, what you write is sensible, but what I think may be hard to understand is how fully consultant-controlled the situation is in the Anglosphere, and how totally consultant-serving the dire USA situation.
In the USA, it is the consultancies who
* choose the projects;
* effectively choose the public agency policy to undertake it (local public officials are incredibly cheap to buy, and public agency staff are unskilled, corrupt and part of a public sector that is expected to fail to deliver and in which professional success is to revolving-door out into the contracting sector);
* undertake “studies” that eliminate project altrnatives;
* actibvely lobby for and receive project-earmarked project funding from larger (county, regional, state, federal) sources (again, political policy is very cheap to buy);
* do all the project preliminary “engineering” with incredible “soft costs”;
* write the project bid documents (effectively choosing the designated construction “winners” at this point);
* be awarded the construction contracts;
* and then game the change order and “unexpected conditions” budget inflation business from both ends (as both the construction contractor and as either “owner’s representative” contractors or as via “public” sector actors who are rewarded for larger project spends)
It’s not that there isn’t state sector geothectnical expertise, it’s that projects are chosen by the private consultant sector to have maximum tunnelling, maximum excavation, maximum site impacts, maximum delivery risk. The chosen projects deliver, and the cycle continues.
It really is self-serving from start to end, with the sole role of the public sector as rubber-stamp for funding allocation. There is no public agency competency at all, and the industry is designed that way, and is flourishing. Never think of any medium scale or larger US public transportation project as a transportation project — it is a funding delivery mechanism, chosen entirely because ot its budget expansion potential.
I’m sorry Richard, but I just don’t buy this story of a nefarious consulting conspiracy. Hanlon’s razor tells me low cost-efficiency in the Anglosphere is more likely to stem from:
– Anglosphere style vetocracy, where lots of small representative bodies and organizations (e.g. utilities) have the power to sink a project at no cost of their own. This breeds defensive engineering.
– Unstable and external funding. It’s easier to skip cost control if it’s not your money. This is more a problem at the representative level, since a consultant will always be yelled at for cost overruns, while a local state rep will not have to face the state/feds.
– Related to the previous, the cost of applying for funding also favors massive projects, which means that funding applications become more onerous over time. Since it’s “free” money from the agency’s point of view, expensive megaprojects crowd out smaller, incremental improvements.
– Heavy reliance on external funding also favours what Flyvbjerg terms “strategic misrepresentation” i.e. fudging the numbers until the they look better than the competition. The pressure to do this usually doesn’t come from the consultant, but the customer, who wants their project shortlisted, the consultants’ integrity be damned.
– Since funding is sporadic, there are low incentives to maintain project management skills.
Case in point, there was an attempt to turn the Finnish Transport Agency into a purely outsourced organization. The FTA even outsourced maintaining the databases that tell what the agency owns and what state of repair their property is in.
If the problem were a collusion between consultants and contractors, this should have resulted in massive megaprojects with equally massive overruns, since consultants were handed the keys of the kingdom. That didn’t happen, because the decision-making environment was and is very different from the Anglosphere. It just led to a lot of projects stalling because of the FTA’s indecisiveness.
In the US, the consultants usually propose overbuilding. The pattern is that the politicians and their appointees know nothing and yet constantly micromanage. In the MBTA example in the post, the RFI didn’t bother specifying an exact loading gauge because the MBTA wasn’t aware this was important (and at TM we didn’t push it because we’d compared the loading gauge with standard European EMUs and it was fine), and now they’re talking about modifying the catenary for a KISS when the euro-KISS fits the current loading gauge easily. Instead of knowing what to specify, Americans constantly ask for small changes, and design-build was supposed to prevent this problem but has made it worse.
The other example I go back to is what someone working on NEC Future told me. The design was contracted out to a consultant, I think AECOM but I don’t remember (could be Arup instead?). There was informal pressure from senior management not to touch anything in Fairfield County, because people there didn’t like disturbances. Thus, the design left it as a magic asterisk, and also included NIMBY-bypassing tunnels elsewhere, because the consultant gets paid based on whether the design is uncontroversial, not based on whether it’s cost-effective.
It gets worse when the consultants are agency retirees, because then all the usual problems of consultant-driven process apply, but also the consultant has the same groupthink as the agency rather than any fresh perspective.
Well I live in Northern California and I know as a direct unambigous fact that every single rail project of the last 40 years was conceived by private engineering/construction contractors, lobbied for by by private engineering/construction contractors, analyzed by private engineering/construction contractors, electoral/political approval bought by private engineering/construction contractors, designed by by private engineering/construction contractors, construction managed by private engineering/construction contractors, constructed by by private engineering/construction contractors.
And when I say “private engineering/construction contractors” I mean pretty much the same couple of them, every single time. The public agencies — whether BARTD, CHSRA, SF Muni, Santa Clara VTA, Caltrain, whatever — are simply host bodies for the consultant parasites. The projects are handed to them from outside and they make no meaningful contribution at all. “Here’s your next project, and who doesn’t like projects?” This really is the way it works.
It’s nice that things function in better ways in other places. It really is. I wasted 25 years of my life thinking that anybody here cared about “better ways” and they just needed to make slightly more informed decisions based on slightly better information. They don’t, and they will not.
I’m sorry to hear that, but it still sounds very much like you need less consultants on that side of the pond, rather than no consultants.
<blockquote.I’m sorry to hear that, but it still sounds very much like you need less consultants on that side of the pond, rather than no consultants.
Patrick, I believe there should be far more consultants!
The existing crowd can continue to try to be private sector consultants, but they’d never win any business based on their well-documented histories of unprofessionalism, lack of skill, and failures of both project estimation and project delivery.
The problem isn’t that there is non-government external involvement in public works here, it’s that a tight cabal of protected (protected from competition, protected from prosecution, protected from oversight) contractors totally control every aspect of major “public” works projects, from conception (“What’s good for General Motors is good for America”) to construction. There’s is no competition and the “market” is closed to exclude competent professional organizations and individuals.
Only works if you then hire someone else to replace them.
I’m starting to think there is some rule like if you do something in-house there should be at least 3 full time people. That way when one leaves you can train a replacement. If you only need one person hire a consultant as needed.
Though you might be able to get some traction by hiring a consultant 25% long term. Jarrett Walker might be willing to do that type of thing: one week a month look at data and make recommendations. The idea here is a consultant that gets to know you instead of coming in trying to learn and make a recommendation before leaving again. Even if you do everything in-house you should get a second opinion once in a while just to make sure you aren’t suck in a bad rut.