Good Practices for State Planning and Local Public Transportation
Earlier this week, I complained about the OPM (other people’s money) problem: federal funding of American public transportation, which is managed locally, leads to cost-raising behavior as local and state governments seek to maximize federal infusion of cash. This is a companion post about more positive and fruitful interactions of government at different levels on this side of the Pond. The examples here often look pointless or acrimonious by local standards, but at the end of the day, they produce cost-effective infrastructure and are positive examples to learn from.
Of note, all the examples below are from unitary, not federal states. This is just an artifact of where I have talked to the most people about this – from what I know of Germany, Austria, Switzerland, and Belgium, they all fall within the spectrum spanned by Italy, Turkey, France, and Sweden when it comes to state-local funding allocation. Moreover, the extent of subnational fiscal autonomy in Germany is not greater than that of Sweden, where there are extensive county and municipal taxes funding subnational government, whereas in Germany nearly all taxes are federal and the Länder mostly rely on transfers.
This is a theme I’ve been investigating ever since I talked to a planner at DOTr. Philippine construction costs are high, although that’s mainly for subways, while elevated lines have fairly average costs. The planner explained to me how planning and procurement are done and specifically how it contrasts with the role of the federal government in the US. Manila Metro projects are planned and designed by DOTr, and ever since that conversation I’ve learned to interpret interviews with European experts in that light.
Sweden: state-local negotiation
The Nordic states practice consensus government. This means that decisions are done by majority vote without veto points, but also there’s no such thing as a majority. In practice, infrastructure involves negotiations between different stakeholders. Bigger projects, including the subway megaprojects we study, require funding from different sources, creating more stakeholders in the process.
In the case of Stockholm, it’s instructive to compare Citybanan and Nya Tunnelbanan. Citybanan is a regional rail tunnel, and therefore the lead agency was the state’s Trafikverket – but even then, Stockholm County had extensive input. Regions send wishlists to the state, and compete for a fixed pot of funding for grants, but there are further negotiations about project details. Nya Tunnelbanan is a subway project led by the county’s SL, but funding comes 25% from local sources, 25% from the county, and 50% from the state.
Crucially, Trafikverket builds rather than just nudges. It has a strong professional civil service capable of designing and supervising the construction of infrastructure megaprojects – and the same pool of civil servants move between agencies within the Swedish public sector, so that some of the people I’ve spoken to have moved between Trafikverket and SL. The example planners I have in mind are mid-level, not top management – this is not a case of a mobile executive suite lording over mid- and low-level career bureaucrats who can’t move between agencies easily.
There is also integration of transport and housing, in the sense that residential upzoning in Stockholm County focuses on areas that have or will soon have urban rail access. Construction rates in Stockholm County are some of the highest in Europe: per SCB, annual completions were around 6.5-7 per 1,000 people in the five years before corona. I’ve been told that it’s a consensual process, with no further elaboration; in Oslo, in contrast, the state has to compel wealthy NIMBY municipalities to upzone as a precondition of giving them subway expansion, but state-local coordination is as far as I can tell otherwise similar to the situation in Stockholm.
Turkey: state-local competition, but no OPM
Turkey has one of the world’s lowest construction cost levels; more details will be available in a report to appear soon, led by Elif Ensari. Wages in Turkey are low by European standards and social protections are weak, but the direct labor share of subway construction is small enough that it is a secondary contributor to the low costs; Turkey dos some things more efficiently than Sweden and others less efficiently.
The situation of state-local relations there is the exact opposite of Sweden’s. There is no collaboration – rather, there are metro tunnels in Istanbul funded and built by the state and others funded and built by the city.
The city is not quite local – the municipality covers the entire metropolitan area of 15.5 million people, and Istanbul politics has an ideological left (i.e. anti-Erdoğan) vs. right (i.e. pro-Erdoğan) characteristic rather than the hyperlocal ties of New York and other American cities. Moreover, now that AKP lost the municipal election and the mayor is CHP’s Ekrem İmamoğlu, who will likely challenge Erdoğan in the 2023 presidential election, there is friction between the state and the city, each trying to argue that it builds more and better infrastructure. There are arguments between pro- and anti-Erdoğan sources over who is to blame, but the city has much less access to state financing now than before İmamoğlu’s victory, which it has been able to replace with financing from the European Investment Bank and other sources of loans, like JICA and Deutsche Bank.
In this situation, there is no coordination, and this is a drag on efficiency – one of the ways Istanbul has been able to keep costs down is finding parks and state land to use for station footprint to keep station construction costs down. However, because there is direct responsibility for the state or the city for infrastructure, there is no OPM problem – İmamoğlu’s political career depends in part on his ability to build infrastructure, and Erdoğan’s ability to interfere is real but limited.
Housing construction is extremely rapid. Istanbul has a housing surplus thanks to the construction of around 160,000 annual housing units; neighborhood character is not a priority there. But I do not know whether it is integrated with subway construction as in Sweden.
France: the capital is the state
France has a convoluted set of local and regional governing mechanisms. However, in Paris, much of the power remains in state and state-appointed organs. The transport association Ile-de-France Mobilités, which would be called a Verkehrsverbund in Germany, is coordinated by the Ile-de-France region, but its two largest components, SNCF and RATP, are both state-owned (though SNCF-RATP agency turf battles remain). Public services that elsewhere in France might be devolved are in Paris often run by the state – for example, the Paris Police Prefecture is part of the National Police, and it’s smaller cities, for example in the Riviera, that have local police departments.
This is not unique to France. In infrastructure, Sweden too exhibits more state involvement in urban rail planning in the capital than in smaller cities – Västlänken in Gothenburg is a Trafikverket project but more of the planning and funding come from the county than was the case for Citybanan. London is a mix: TfL is run by the mayor, offering much more devolution than the Metropolitan Counties of England have, but conversely the construction of infrastructure megaprojects like Crossrail is really within the purview of UK-wide politics.
The issue here is one of scale. Grand Paris Express is a 200 km, 80% tunneled project, and France is a medium- rather than low-cost country. Even the state barely has enough planning capacity for it – the Cour des Comptes report on the cost overruns, not seen before for smaller Métro extensions, blamed the insufficient size of existing planning organs, but unfortunately, the solution arrived at, the special-purpose delivery vehicle (SPDV) GPE, is not good, and is either in imitation of or evolved toward convergence with Crossrail. Nothing below the level of the state could build such a project.
And because the project is so large, it’s been forced into a situation that rhymes with Sweden’s intergovernmental negotiation. It’s also been discussed as part of national politics, with some redesigns stemming from the Sarkozy-Hollande transition. In some cases, this has led to OPM – namely, M18 is unpopular among the region’s public transportation advocates and remains because of pressure by the high-income suburbs it would serve. However, there is no visible impact on unit costs; it’s notable that the OPM the state would dispense is additional infrastructure at per-rider costs that are high for France but common in the United States, rather than extras of little use like signature stations or more expensive construction methods.
Finally, housing construction in Ile-de-France is, as in Stockholm County, among the YIMBYest in Europe. Yonah Freemark’s paper on the subject is indispensable: stating around 2017, the annual construction rate rose to 80,000 units regionwide, around 6.5/1,000 people. Construction is largely in the Petite Couronne suburbs, and not the city, and focuses on regions with current or future urban rail extensions, as in Stockholm.
Italy: state planning and austerity
A full report on Italy will appear soon, on a similar timeline as Turkey, written by Marco Chitti. In Italy, there has been a transition from municipal funding and planning of metros to state funding; in Rome, there was always more state involvement as I understand it.
The situation leading up to the Financial Crisis had similarities with the United States: state funding, municipal or regional responsibility for construction. However, the state always exercised far more oversight. The Italian state builds rather than just nudging. State regulation is done through administrative rather than judicial mechanisms, and thus questions of environmental and historical protection are decided by civil servants trained in engineering, archeology, history, and ecology; there are clear rules, providing similar final outcomes to the Nordic process of negotiation and superior ones to the American process of lawsuit.
More recently, the state has devolved some of the funding to regional, provincial, and municipal governance. This was an artifact of post-Crisis austerity, so the state would fund the majority (I believe 70%) of each project’s budget but not all of it. The result has not been positive – subnational governments have no money, not even wealthy ones like Milan, and to fill in for missing state funding they’ve resorted to PPP financing, which has not impacted construction costs but in effect required hidden loans at high interest bonded to future revenue.
Pedestrian Observations 
In the USA, the federal funding organ (FTA or occasionally FRA) has to complete the following:
1. Environmental Review: This is normally outsourced to consultants such as Booz Allen Hamilton, Lockheed etc. If you are lucky, this will take 4 years
2. Design/Engineering Review: This means the feds review the documents prepared by AECOM/WSP/Parsons/etc. for the local agency applying for the grant and sometimes hire (hire meaning spin off a new work order on a massive ID/IQ contract awarded years before by USDOT) an engineering consultancy to do the review because the agency staff have no experience with the system being proposed leading to further delays.
3. “Cost/Benefit Analysis” This is often actually done in-house to my knowledge and because there are no standards specified in legislation, the CBA is not stress-tested by the feds.
Keep in mind that throughout this process, the consultant teams at the local level and federal level usually cannot talk to each other. Everything has to go through the full time federal/local (e.g., MTA) employees. So if the federal environmental consultant needs something, they have to convince the federal employees to raise the issue, then convince them to ask the local agency, and hopefully that local agency still has funding to get something out of the consultants who once paid will often swap qualified resources for inexperienced staff that they pay less but bill the same hourly rate to juice profitability. It’s an absolute mess!
Then, if the project is delayed 2 years by the environmental process, funding problems, political meddling, etc. The whole team has fallen apart by then, no one can remember where the original .ai, .docx, .csv, .xlsx etc. files are and some work has to be repeated!
The feds disburse money once a process is completed. It’s about following process to check boxes, not about delivery quality, cost control, or delivering modern, efficient transportation systems.
I should add that most employees on the Federal and Agency side know that this process could be much more efficient. But they are hamstrung by legislation, and the potential for lawsuits if the process is not followed exactly. This leads to a ratcheting-upwards of conservatism: “well we put this in the EIS last time and we have this additional data, so let’s throw that in too, can’t be too careful!”
EISs used to be 80 pages long and they have grown tenfold!
I’ve just scanned Yonah Freemark’s paper. The essence appears to be: ” …providing additional permanently affordable units— counters declining per-capita government support for subsidised units, a trend in countries such as the UK and US … and addresses lower income family needs.”
That is, a big increase in social housing that France, like the rest of the rich world, had stopped building as high post-war growth rates declined in the 70s-80s. Fortunately not in the high-rise concrete horrorshows from the earlier social housing era. It’s kind of what many have been saying for a long time, and lo, when a region reverses such passive policies the housing affordability issue abates (in the Île de France region which had these policies; as YF points out, this was not nation-wide policy). An interesting thing is that he ascribes the turnaround to a combination of Paris Mayor Bertrand Delanöe and President Sarkozy, particularly the latter’s Grand Paris scheme. There’s the usual role of TOD, ie. promoting densification around new metro (RER) stations, and using various state instruments to nudge local authorities who tend to avoid densification if left to themselves.
I’ve usually minimised Sarkozy’s role in this, assuming these things were long brewed within the French deep state, but perhaps I should overcome my biases. Not sure. In any case Sarkozy’s motivation was driven by higher economic development concerns that simply overlap with the likes of Delanöe’s (and doubtless his deputy, Anne Hidalgo’s) social equity concerns. This has been stressed by Nobel-prize winning economists around the world since the GFC but largely ignored by right-wing politicians. I’m wondering if the passing of SB8 & SB10 by Gavin Newsome shortly after his win in the recall election is the beginning of a real change in the US, led as usual by California?
Incidentally I was having this exact argument with someone at my usual Saturday morning cafe and my companion brought up on screen a list of “average housing cost as a multiple of average income” and these things show what appear anomalies. Certainly I cannot explain things like this:
Country………………Ho.Price-to-Income Ratio……Mortgage as %Income
Hong Kong………………..45.71……………………………289.30
South Korea………………26.08……………………………178.71
Singapore………………….17.63……………………………108.34
Japan………………………..11.59……………………………..66.64
France……………………….10.04…………………………….58.13
UK………………………………9.46…………………………….63.66
Germany……………………..8.88……………………………..53.09
Switzerland………………….8.18……………………………..47.17
Sweden………………………7.71………………………………49.51
Netherlands………………..7.28………………………………44.50
Australia…………………….7.27……………………………….51.48
Canada………………………7.24………………………………49.62
United States………………4.04………………………………29.42
As expected HK is top of the list but I can’t see why France would be ahead of the UK (or Australia). I also don’t really understand the second column (which is I believe the average mortgage repayment as percent average or median income). It has to be some sort of averaging artefact because I am pretty sure France (and Germany and others on this list though not the worst like HK, or the Anglosphere) has strong prudential controls on what banks can lend you, and it is way below allowing mortgage repayments eating 58% of your income! Is it that this totally ignores social housing? Which could explain why France & Singapore are higher on the list than expected while still housing most of their population “affordably”? So the people buying houses that contribute statistics to such a table are from the higher income group and of course their percent of their income is way lower?
This is a hot political issue in Australia because housing has increased by >10% in just one year and is clearly unsustainable. Of course this is what happens with close-to-zero interest rates. New prudential rules on banks balance of first-time-buyers versus so-called ‘investors’ are coming in, but politicians are always very nervous about any messing with the housing Ponzi scheme crashing the economy …
I wonder what today’s vote in Berlin on effectively nationalising the big housing companies will be? Or if it can possibly be put into effect.
You have to be careful with mortgage as % of income, since it doesn’t tell you about the size of deposits or length of time spent paying off the interest. Low interest rates make large mortgages affordable on a year by year basis but if you spend an extra decade paying it off that’s a very different ballgame despite the same interest rate. And rents can be quite different too. And then there are social housing rates (the UK is still the highest in Europe at nearly 25%).
Essentially the banks used low interest rates to extend leverage to homeowners in capital-open-high-building-restriction anglosphere countries. When the building restrictions of Postwar regime start to bite in the 1980s while social housing construction tailed off, homeowner Conservative coalitions headed crisis off with financialization instead of building reforms. Without the banks we would have had a rent crisis in the late 1990’s*, instead housing bubbles. The hilarious thing is the banks are the least corrupt people involved! They expanded output and cut prices!
*Germany at the present is example of what this might have looked like with the Berlin nonsense and Scholz’s building promises
The German problem to me seems to be that local and state governments have a lot of ways to kill projects the feds want but don’t have the converse power of forcing projects the feds don’t want to go ahead. Look, for example, are the Cologne-Frankfurt HSL which had to include stops at Montabaur and Limburg to overcome opposition of the state governments of Rhineland Palatinate and Hesse…
Of course with tram projects it’s often not even local governments which derail (pardon the pun) them but referenda which – despite often running against something like 80-90% of seats in the municipal council – are often taken as gospel even if they only represent a 55-45 with 25% turnout… And that even tho the binding period of referenda is usually only a year…
In Switzerland (which has more referenda and a longer tradition of them) a “lost” referendum for a public transit project is usually seen by its boosters as grounds to retool the project (cut costs, increase scope, etc.) And try again a few years later…
A further problem is despite the (semi-) objective criteria often preventing quality degradation or scope creep (to take the http://www.stadtumlandbahn.de as an example, the proposal to take an overly circuitous route to “protect” a river meadow already compromised by the A73 highway fails to receive a benefit cost quotient above 1 and thus won’t be built), they also fail to account for many important things and undervalue others. For example, many very desirable capacity increases have no chance of getting above 1 due to capacity and stability of schedule not being valued and instead often being deemed “waste” as the schedule assumes *best* and not real world conditions…
To add a further example from the above mentioned case, a rerouting nearer to the main lecture hall (German/Latin term “Audimax”) to be while also serving the main student cafeteria (German/Latin term “Mensa”) leads to a lower benefit cost quotient, because the assumptions do not properly account for university students (the assumptions seem to calculate ridership based on housing units and/or jobs first and foremost) and the perhaps one minute of longer time the rerouting takes (as opposed to a “straight shot” down the semi-pedestrianized Nürnberger Straße) is weighted more heavily than potential ridership gains…
To say nothing of highways still getting federal money for some reason passing understanding…
BTW, Alon, if you want to make this knowledge easier to understand for us ignorant Americans, maybe you should use the term ‘national’ instead of ‘state.’ I know that the way we do things, referring to reigons as ‘states,’ doesn’t make any sense, but part of my brain still automatically applies it to other countries.
Indeed. The word “state” has too many meanings not to be confusing — for instance also “government” in general, on any level.