Institutional Issues: Professional Oversight
Continuing my series on institutional issues concerning infrastructure costs and quality, after the issue of procurement, I’d like to discuss the issue of the quality of public-sector oversight. It is critical to have extensive in-house expertise inside an apolitical civil service empowered to make technical decisions. The role of the political layer is to set up broad rules, not to micromanage. Conversely, while the top people should avoid micromanagement, they should be expert enough to be capable of making specific decisions.
Civil service and oversight
The importance of civil service to oversight is that it’s the professional layer that has to supervise planners, engineers, architects, and construction teams. There are too many small decisions for a single elected political layer, say a minister and a policy team the minister directly appoints and supervises.
In my procurement post, I was basing my recommendations on common threads I’d seen or read about in low- and medium-cost European countries, and to some extent practice in South Korea, a low-cost country on a par with Southern Europe. All of these make use of professional civil servants to make any of the following decisions:
- In-house planning. The macro-level decisions on funding levels are political (and never devolved to the agency through dedicated slush funds, unlike for US highways), but the decisions at the level just below, such as what programs to ask the politicians to fund, are made by professional agencies. High-speed rail was invented this way in Japan and then reinvented in France, while upgraded legacy rail was so invented in Germany and perfected in Switzerland.
- Design and engineering. Those can be done in-house or outsourced to consultants, or more likely some mixture of the two, but even if the design is contracted out, it’s the agency that owns the product and is responsible for it.
- Contractor selection. It is irresponsible to award a megaproject design contract based on the lowest bid. A technical score is used nearly universally in the low- and medium-cost examples we have looked at, and this means someone needs to come up with sound criteria for scoring and then evaluate each proposal. This has to be done more intelligently than just by rubric. A British source told us of a problem with British technical scoring: every large project is parceled out between different consultants, and thus all consultants can claim experience on the same project, making it impossible at that level to tell which companies do better work than others, even as industry insiders know who does bad work. The same source, when I asked about French comparisons, said that France has extensive in-house expertise and therefore doesn’t hire their consultancy.
- Contract supervision. Change orders are inevitable, especially for underground projects. Not coincidentally, in Eno’s database the American premium for subways is higher than for at-grade light rail, which is technically more predictable. It’s on the client agency to decide whether to accept or reject changes coming from unpredictable factors, and this requires extensive knowledge of the field.
In my procurement post, I spoke of flexibility. No client can have flexibility without oversight – flexibility without oversight is an anything-goes game in which the contractor abuses the client, the client abuses the contractor, or, most likely, both. And this oversight is necessarily detailed enough that it requires civil service.
An example from Sweden
I spoke to an experienced Swedish project manager earlier this year. The project manager talked to me about the major issues with the construction of Citybanan, the regional rail tunnel Stockholm opened 4 years ago, shortly after I left the country. This included issues of construction techniques (but for further engineering question my source referred me to an engineer) but also competition policy, negotiation, change orders, etc.
At one point in the interview, I asked about something a previous Swedish source told me about, called functional procurement. In functional procurement, the agency maximizes flexibility by specifying only the function of the project, such as the required capacity and schedule, and letting the contractors make suggestions as to how to fulfill it; this is similar to the military concept of mission command, stressing flexibility and training intermediate officers in how to use it in a hierarchical organization.
The project manager said of this growing trend in Swedish procurement: “I can’t say it makes it easier.” The manager then explained the constraints involved – railways have technical specs that make a functional contract not too different from a conventional one, where the design is already figured out and the contractors have to build to it with only minor modifications.
Let’s unpack what happened in that interview. A Swedish manager who does not know me, who I have never met, first of all talked to me in technically adept language, and second of all was willing to go on the record criticizing a trend in infrastructure construction procurement, a trend that the person who put us in contact had mentioned to me as a positive.
I have never heard this kind of internal criticism from American sources, unless they knew me well enough and were trying to get me to publish their internal problems in the media. And quite often, the criticism I would hear from the US was much more pungent, complaining about politics or a bad manager. “We have been trying this trend but I don’t think it’s working,” in exactly the tone you can imagine emerge from the style of quotation, is not a line I recall hearing from an American. The civil servants who criticize something are far more frantic, far more afraid. Sweden will trust its civil servants to literal death. The US (and UK) will not trust them to do anything but follow orders.
Is Sweden unique?
No. Strong traditions of professional civil service exist everywhere we’ve looked outside the US. Even the UK has a semblance of it; the problem there is that the topmost layer of civil servants – the Sir Humphreys – consists of lifelong generalist elites rather than domain experts.
In Italy, the situation is especially lopsided. The political layer of the government is weak because party control changes so often and ministers do not last, and there are so few political appointments that even with political instability, the civil service lasts across those changes. If anything, the instability makes the professional layer stronger.
Apolitical experts
It’s critical to ensure the civil service is not political. This doesn’t just mean that it should not be partisan. There are enough dominant-party jurisdictions in which it’s understood that the civil service exists to implement a predictable political agenda, which can be left-wing (Berlin, New York, California) or right-wing (Bavaria, Texas). Those jurisdictions all have problems coming from the lack of meaningful political competition, but those problems come from politicians, not civil servants. No: political noninterference goes much deeper, and means sidelining issues of petty personal priorities.
The ideal civil service has as few political appointees as possible. Those are neither elected nor meritorious. By their nature, they lack the legitimacy of both the politicians above them and the deep layer of domain experts below them. If they’re selected from among people with industry and operating experience then this is fine – technically senior generals are political appointments in both the US and Israel subject to the usual military norms, and Andy Byford was an external hire for New York City Transit with experience in the industry but not the agency. But letting generalist managers selected for political loyalty parachute in charge of agencies is a recipe for disaster.
The word for people who are always to be managed by people who are not from within their own social group is servants. Such people, knowing that their manager will always be someone who has other priorities that are not always transparent to them, will learn to lower their heads rather than proactively looking for ways to improve their institution.
A scientist working at a federal institution explained it to me this way: “There’s absolutely ways to speed things up, but they need cover from the political appointees at the top. A career officer understands their role to be following an existing plan, laid down in writing by either congress or by a planning process involving the top (i.e. political appointee) officers of the agency.”
This was meant to explain the sluggish FDA corona vaccine approval process, but can equally apply to infrastructure and operations of public transportation. Any variation from a plan written long ago by people who were often not even at the frontier of knowledge then requires political approval.
Trusting the civil service
A low-trust society isn’t one in which common people don’t trust the elites. It’s one in which the elites don’t trust the common people. In the context of civil service, it’s crucial to set up a system in which the top people affirm rather than scourge those below them.
Byford did it well, setting up a system encouraging employees to complain and suggest improvements, much to the surprise of managers at other MTA agencies who preferred scourging their subordinates. At the topmost level, it means the political layer needs to affirm the authority and expertise of the civil servants; in conflict between a petty actor such as a community advocate and the junior members of the state, the state must support its own, while internally ensuring that the proposal has technical merit. (Political merit is judged by elections, not by who screams the most loudly at midday community meetings.)
Civil servants who see that their superiors are hired and promoted from within the ranks or among peers, and judged for their ability to work with those below them and not just those above them (in the tech industry, a managerial hire spends some interview time with the team they are to supervise), will soon learn that they can show initiative. The ones with bad intuition will fail, whereas the ones whose initiative is more successful will be able to rise and transmit their ideas to the group. It goes without saying that this also requires staffing up to normal levels and paying competitive wages. This way, the state can ensure it can oversee its own projects competently; there is no alternative.
Pedestrian Observations 
This is a great article pointing out THE major flaw to all of our Civil Service systems, and why “public works” in this country is so inefficient and costly. I would add another layer of WHY this is so…
I believe that in some other advanced countries there are actually specific educational institutions that prepare Civil Servants to serve in their respective professional capacities. This provides an institutionalized level of purpose and professionalism. It also supports a broader and shared cultural understanding of the importance of persons to serve in these capacities. Which more naturally includes an appreciation of the importance of maintaining healthy ‘boundaries’ and practices that avoid political distractions and the manipulations of powerful business and labor sectors.
Without a robust Civil Service ‘Infrastructure’, we’re often left with a wide variety of inappropriate, undisciplined -and therefore rightly under-appreciated- civil management, very vulnerable to political confusion and manipulation.
As there seems to be no mention of this in the discourse around the current Infrastructure Investment Act, I fear we’ll just continue to bumble forward as usual, and waste both the money and opportunity to improve the system, achieve objectives, and do everything better from here forward.
(-please someone show me otherwise.)
France seems to have such specific educational institutions, whereas Switzerland does not. As such institutions tend to be very state specific, and discourages ideas from abroad. French railways are extremely averse at any foreign ideas, and this went to farcical level on the CEVA / Léman Express project in Geneva…
CPOC (pronounced see-poc) the Capital Program Oversight Committee at the MTA, is getting its name changed to Capital Program Committee, it was announced (well not really announced but somewhere in the fine print) this month. “Vee Know NUSSINGK!” …along with a new chairman: one of J. Lieber’s friends from the Dept. of Demise and Destruction.
Not that I know anything about him, and Lieber seems off to a very good start. I DO know that concerning one project – to change the Keener Shelter building on Wards Island (a former tuberculosis hospital) over from city steam to local boilers, they made quite a show of it in 2014 – flying the boilers into their prepared-at-huge-cost-of-time-and-expense positions by giant construction crane. It was really thrilling.
The city steam in this case came from the nearby State Psychiatric Hospital, technically on Randall’s Island though both islands have been fused for quite a while now. I’d say the old boilers were about 2250 feet from Keener, which was on a circuit of four or five large buildings, including the giant Psych Hospital and the water treatment offices, that got their winter heat – very efficiently – from the old steam system. Though it was constrained to a degree by narrow underground steel pipes of less than 6″ diameter, which sort of lost some heat in the steam transporting process I’m told. NOT AS INEFFICIENT AS THIS CONSTRUCTION JOB THOUGH.
Attached to the flown-in boilers are lengths of 3-foot-diameter steam pipes in double configuration, outbound and inbound, positioned in wide ditches 7.5 feet wide by 15 feet deep in stacked configuration. Because the boilers were mislocated – ostensibly for aesthetic reasons, to prevent blocking the view of Keener’s front elevation from the soccer field there in the shadow of the Hell Gate Bridge – they required a greater-than-normal number of expansion loops, in this case three, for each direction of piping, which made the ditches about twice as long as they would be otherwise. (An Expansion loop in this case, apparently, must extend from the straight line of a non-looped steam pipe some twenty feet laterally away from, and about ten parallel to the line, before returning to it, which adds quite a bit of complication and expense as to building it.)
Apparently there wasn’t sufficient oversight at D&C to prevent the additional padding out of this basically questionable project, with like ones around the city involving cutting off the steam years prior to any replacement being available (i.e. built) and requiring temporary “portable boiler services”. (not saying these are in any way connected with anyone at D&C)
Just an example of what we’re faced with in construction and procurement, and the relative hopelessness of achieving any kind of effective formulaic solution. Of course it’s NY, but other locales LEARN from NY. (both as to innovations like freezing the ground for temporary stabilization, and idiocy, like running a further extension of the No. 7 Line to NJ ((for which the ground-freezing method, first used in the previous extension of the same line – which as rendered is practically useless – would be optimal)) If this isn’t the type of problem we’re facing with rail transportation in the US I don’t know what is.
There are, I know, a lot of lawyers at D&D – and this has been the case with transportation companies, particularly railroad companies, and all the others that copied them, going back a long ways. Lawyers and engineers. It occurs to me the engineers see the fiduciary lawyer types as adversarial and are often engaged in schemes to thwart them in their mission – if they aren’t complicit. It’s been going on for a long time, hidden in plain sight.
I tell you I can think of DOZENS of things like this steam pipe scam and far worse just at the MTA – some of them already etched in stone and steel. It is always the tendency to INCREASE both timelines and work to be done – by whatever means necessary – in any construction project you might happen to investigate. Do they have Inspector Generals who are conversant in engineering matters? I think not, or at least not if they aren’t on the take. What is required is FIDUCIARIES WITH DISCERNMENT, and to have them backed by reliable engineering advice. It’s a tough row to hoe.
Has any developed country recently, substantially improved their civil service though? If so how?
The post very convincingly explains why a problem exists, but I worry it also identifies a near-insurmountable roadblock to improvement. At least with overly expensive stations, one could recommend less fancy stations. How does one recommend a more competent civil service?
Alon’s main case of improvement is Italy which was high cost in the 1970-80’s before the big scandals, the collapse of the Christian democrats, the crackdowns on the Mafia led to a reassertion of experts in Italian rail construction and a decline of costs, at least for urban rail. But I guess we are waiting for Marco Chitti’s report on Milan for something detailed in English. They also show that the Boston Green line costs improved after they did a project reshuffle and cut a lot of the fat.
Has anyone ever tried “hire staff who can write and enforce technically scored itemised infrastructure contracts like the Spanish and Koreans” as a slogan? I doubt it, but it might coherent enough to start changing minds at the elite level. It’ll be bespoke to each country though.
@Borners – Yes but Italy, France and Spain have never fallen into the decadence of malevolent engineering malfeasance as we have, and as Middle Europe has flirted with (see Mehdorn et al, and the Semmering joke – remind you of any Hudson proposals?) and as England is about now going off the cliff of HS2. (for an as yet undetermined fee) – I hope they never do.
Its Britain that’s going off the rails with HS2, Scotland has the same cost disease. England’s not allowed to exist, that why the English have treated the British state with contempt since the 1970’s.
With the US, if the introduction of Civil Servants in the first Guilded age was possible, it should be possible to reform it, in core Blue America at least.
I see the problem with HS2 as different, and the whole issue of civil servants not capable of sufficient discernment as something that’s really not their fault, though in practice it is. There is a malevolent strain to railroad engineering in particular. You can see it going back to the 1st part of the 20th Century. The Lackawanna Cutoff opened 1913 was level to the point of ridiculousness, at huge additional cost. I think the ruling grade is like 0.4%. The fill embankments are a scar on the landscape. Of course steam don’t go up hills too good, but no doubt the president of the Lackawanna was expecting electric was coming. (He was very forward-thinking.) The New Haven Line had already been running to Stamford for five years when they started building the Cutoff. The Roseville Tunnel on the same line was something that caused a years-long dispute between the president with good instincts, and the engineers with their malevolence. They said the tunnel being located in an anticline was good for tunneling, rather than a cut. It looks like a syncline to me. The president wanted a cut. I say it should be daylighted if they’re going to use that line again, It’s dangerous. But now the FRA and Amtrak – all bunches of engineers fused at the hip and functioning as a malevolent cabal, say it has historic significance. I think they got it on the register, whereas you can’t register the Hotel Pennsylvania, because they want to build another chimney-shaped superscraper there. Actions have no basis in rational fact.
Anyway the pennsylvania Cutoff, an extension of the one in NJ that runs N. from Scranton, ended up having the wrong route because they wanted to build the giant bridge in Nicholson, that towers over the town and stunted its growth. Made a shitty passenger station as it should have been located IN the town rather than up by the bridge. But they wanted to build the world’s biggest reinforced concrete bridge.
Anyway, with HS2 it would be so much cheaper, easier, less intrusive, better – to build tunnel and other abbreviations, which would be all interconnected to the existing lines and stations – greatly increasing capacity and versatility, with six-tracking where necessary. (It’s a lie that can’t be done. It’s only necessary to measure it in Google Earth to see that.) That way it could have been staged, and taken in bite-sized chunks. They build these giant curves that trains grind around at high speed incessantly – which is actually longer in linear feet than a line that’s been abbreviated properly. It’s very expensive to run trains that way. (See the Hudson tunnel Project for another example. The track will need to be constantly maintained with those curves – yet the old tunnel – if it were sound – would always be a quicker trip. Here we revert 110 years and come up with an inferior solution!) And they have never gone to linear, whereas you can go 300mph with Maglev, and the technology is available for conventional – no catenary or third rail – but they will have none of it. Without the square wheels and skids it would take railroading out of the Dark Ages. THAT’S WHY they will have none of it. Whereas HS 1 was the greatest passenger line at the time, HS 2 will be garbage from the beginning. You can tell by how they start digging up and demolishing everything around Euston what they’re up to. It’s just destructive. A pathology. I haven’t studied it enough to know what a comparable capacity solution would be but there’s no question it can be done cheaper, better, without the upheaval. It is designed to give passenger railroading a bad name.
Well it would seem right to hold the technical experts accountable when they give bad advice. This happens often, but it’s difficult to tell who of the fiduciaries are complicit. In the case of the PANYNJ, capital projects and planning seem all to be filtered through one Executive Director, who is notoriously generous with funds and appears to be giving the planning contractors what they want, in terms of both exorbitant funds and carte blanche as to their preferences with physical particulars of any given project – which are often poorly conceived and inadequate, as to both excessive cost and relative uselessness. Of course the commissioners all vote in unison without fail, and have done so consistently since one of them objected to the retrograde LaGuardia Airtrain in 2014 – which made quite a splash in the news because he was not at all moderate in expressing his disdain for the entire thing. It seems there’s a likely sign something’s wrong when a board like that ALWAYS votes in unison, and they ALL do – NJ Transit, MTA, Gateway Commission, whatever – always. Anyway the commissioner was sacked. (The only one with any sense if you ask me.) So here we are six years hence and they’ve wasted at least $500,000.000 on this idiotic project. (though it’s possible not all of it was spent)
It’s not because the board members – OR Cuomo – prefer one particular mode or physical configuration over another. They take their advise from the professionals and consciously lack the discernment necessary to recognize when it is a mischievous choice – unless they’re stupider than that. Then you get the type that apes the sentence construction in the environmental docs but fails to understand what he’s so worked up about and calling you an idiot for. It’s a tiresome endless circuit. For instance, I’ve tried to warn the River Keepers (they haven’t called me an idiot – this is a different thing, not a for instance) in their suit not to insist on extending the N + W trains because it’s complicated and problematic, AND the Port Authority has carefully formulated their arguments against that option and will surely win if it comes to that. From what I hear of the suit they didn’t listen to me, and I can barely get a response out of them.
I have produced two detailed alternative options for the Airtrain, one of which has been on the internet since 2014, and there is another one I’m holding back on – not really mine – because I want to embarrass the PA concerning their lying about it to get it eliminated from consideration, as they have about both of mine, with not so much as a mention in the final EIS concerning the long-lived one. They’ve been quite provably dishonest about all this, in writing, but that was before Hochul pulled the plug – and it just shows the type of reprehensible engineer flunkies that make up the intelligence pool in this greatest city in the world in 2021. Here’s a link if you want to know more about my two correctly conceived and clearly-thought-out alternatives – which is more than I can say for anything the PA has does infrastructure-wise in about twenty-five years: https://www.rail-nyc-access.com/laguardia-direct-route
I didn’t read it, I looked at the pictures.
If you goal is to get people to LaGuardia Airport there are lots of options. If your goal is to get the most air passengers in and out of metro New York airports, close LaGuardia. Using the LIRR/Amtrak tunnels is not an option. You will be trading packed 8, 10 or 12 car trains for half empty 6 car trains.
Meant to say – or just thought of it actually (this reply comes after the one that appears to be under it at this point) – that’s the way to read environmental documents. (not to attempt to dignify my my techno-screeds by calling them environmental documents) but who can plow through that stuff? …the way they write it, with all the mandated details, ((and 47 alternatives in the case of the AirTrain)) But there’s more to it, and I don’t get the impression that transparency is always the goal. When both the text and the picture is unintelligible, then you’ve got a problem.) But yes, the quickest way to TRY and comprehend the meaning of those things is to hunt through the sidebar for pictures that actually depict what they say they’re planning to build in the text.
There can be problems with that though.(see picture) Here we have this joke labeled Harold Interlocking, but the junction is off the picture to the right. This is the sole illustration of the Amtrak Bypass Tunnels, both of which have their plans re-formulated at least twice. I can’t tell what it means. In other drawings the mess running off the ESA Line split to four leads is called Plaza Interlocking.Probably that doesn’t include the Bypass Tunnels. First they were cut-and-cover, then one was to be bored. I know one was, because I saw them removing the TBM.
Oh wait, I can’t post the picture. Seems to me I’ve posted pictures here before…
Post a link to the pics?
47 alternatives in the case of the AirTrain
If they don’t include a paragraph on why dirigibles to the mast of the Chrysler Building isn’t a viable alternative someone will sue because they didn’t examine it.
Well you might be surprised. There are a lot of constraining issues, obstructions, with getting to LaGuardia, but it doesn’t justify doing it as backwards as possible. I don’t get what you mean about closing LaGuardia. You could get fast service and more volume to Kennedy on the Rockaway Beach Branch. You’d have to run short cars into Penn though and there’s no room in Penn. Except for Penn Sta. Access heaven help us.