Privatization is not an Alternative to the State
There’s a set of norms that are required for successful governance in a developed state: transparency, professional civil service with no political overclass, strong political parties, consensus that the basic functions of the state should exist. The state cannot provide public services like infrastructure, health, education, and security without these. States that lack these try to engage in workarounds, like privatization. In infrastructure, this is where the globalized system comes in: large in-house bureaucracies are dismantled or not built in the first place, and in their stead are multi-billion-dollar contracts to consultants, overseen by political staff. The system of privatization sells itself as an alternative to the state if there’s no consensus, but it isn’t one. The result is high construction costs for infrastructure and such long timelines that it’s impossible to build anything.
As we point out in the Swedish case, the distinction between the traditional and globalized systems of infrastructure project delivery can be roughly summarized as the following:
| Traditional system | Globalized system |
| Design-bid-build (separate contracts for design and construction) | Design-build (one contract for both design and construction) |
| Smaller contracts, typically tens to low hundreds of millions of dollars | Larger contracts, starting in the mid-hundreds of millions of dollars (Sweden) and going up to the billions (New York) |
| Itemized contracts with publicly available prices | Fixed-price/lump-sum contracts |
| Product procurement: the contracts list what there is to be built | Functional procurement: the contracts list what the project is supposed to do, e.g. capacity, leaving the mode underspecified |
| Public-sector risk | Private contractor risk |
This, to be clear, needs to be nuanced. The most important nuance is that despite the name, the traditional system is better viewed as a modern evolution of how contracts done in the first three quarters of the 20th century, with an eye for far better anti-corruption mechanisms. In particular, itemization is a fairly new innovation – in Italy, item-level transparency is a 1990s innovation in the wake of mani pulite.
What’s more, quite a lot of systems are mixed. The Nordic countries have been moving toward the globalized system on the theory that the private sector is more efficient than the public sector, and their construction costs have exploded in tandem. But their moves in this direction remain more halting than those of the United Kingdom or Canada, or as of late the United States; the contract sizing is well below American norms, and, at least in Sweden, Trafikverket makes sure to retain in-house capacity to oversee the contractors and consultants better. In contrast, the United States has a system of consultants supervising other consultants.
And finally, as is usually the case with privatization, the more private contractor-centric globalized system does not mean the state actually retreats from infrastructure. To the contrary, the political elements of the state have greater control. The staff of an elected mayor, governor, or prime minister are more involved in $2 billion contracts than in $50 million ones, and can influence the bidding based on criteria that are never how to do good work (they’re political staff, they don’t know what good infrastructure planning is and don’t respect people who do). This is not privatization in the sense that 19th-century railroad construction in the United Kingdom was financed and designed by entirely private actors, with the public role restricted to parliamentary approval of routes.
It just so happens that none of this works. In Sweden, there’s enough state capacity that people are questioning the globalized system, saying it is designed around the needs of multinationals rather than those of the Swedish public. The long-term attrition under which British and American civil service bureaucracies have atrophied has not happened, so Trafikverket could take a more active role if it wanted. But in Britain and the United States, it is harder, and professional bureaucracies, run by engineers and planners and not by political appointees, need to be built.
If the state can’t build, then it can cut taxes, substituting private consumption for public infrastructure. This is fine; Americans have large cars (often to protect from other large cars) and large houses. But it’s not a substitute for infrastructure and other social goods; it’s a conscious decision to lose years of life expectancy and have a less efficient transport system to avoid building up the state.
Pedestrian Observations 
Consultantism (esp. in its Anglophone form) sounds a bit like a parallel to feudalism. Another answer to the question of how to govern without bureaucrats — it’s just that this time, it’s not because they don’t exist, but because your cultureligion has forbidden them.
Its not really. The late Roman civil bureaucracy collapses because the successor states in the West don’t adopt the tax-paid army model, and the economy collapses at the same time in the violence and the break-up of the Roman trade-zone.
Closer is what happens in the Islamic world with the Iqta (which in turn becomes the Timariot/Zamindar/Pronoia elsewhere) where the collapse of the Caliphate leads to a Turkish military caste weirdly refusing to become a local aristocracy but instead an effigy of a state. Although even there that system is a symptom of the wider Islamic world outgrowing the Caliphate and the collapse of Iraqi agriculture over the 800.
And the entire High and Late Middles Ages is about rebuilding state capacity, with the Church being a leading part of that system (Canon law informing secular law, providing the personnel etc). The reason it takes so long to move a “modern bureaucracy” is that in a agrarian society with low surplus its actually not necessarily cost-efficient. Indeed the Reconquista in Iberia constantly has the Christian “feudal” states outperform the more “bureaucratic” Andalusian states. And if the Ottoman’s hadn’t inherited the Byzantine fiscal infrastructure they would never have penetrated Hungary (still needed a lucky Break at Mohacs)
And even supposedly the most “bureaucratic” major culture China, scratch below the surface and rural late Imperial China is often more anarchic than late Medieval Europe because just sending one guy with a degree to rule 500,000 people with no paid staff and no army does not a bureaucratic state make. Jinshi grads were closer to Early Modern English JPs than modern CCP magistrates or contemporary professional bureaucrats.
Thatcherism/Reaganism are specific outcomes of distinct legitimacy crises of both states. The former is about the failure of the British national projects of 1922-1979. The latter, the breakdown of the Post-Reconstruction social consensus. Anti-state statism, emerges when you confront social changes you can’t stop, good governance will intensify, and when the state is your strongest weapon (in the UK/US social conservatism is stronger in government than anywhere else).
I wasn’t predicting the consultantification of taxation, the military (PMCs), or the police.
However, Alon calls consultantism “the global system” exactly because it has spread beyond the US&UK, into countries which have not had those two specific circumstances. Some of the selling points are obvious:
1) Politicians are afraid of the bad PR of cost overruns and suchlike. They are perfectly willing to overpay by tens of percent for the prime contractor to take the bad-PR risk (while the politician still takes any good-PR). This mechanically requires for there to be a prime contractor.
2) “To the contrary, the political elements of the state have greater control. The staff of an elected mayor, governor, or prime minister are more involved in $2 billion contracts than in $50 million ones, and can influence the bidding.” One person’s drawback is another man’s benefit. Finally I can spend the tax my constituents pay on buying said constituents’ votes by holding community meetings and scoping niceties into the project. Is this more efficient, even at the goal of buying votes, than building no-frills functional infrastructure&service? Dunno, but this consultant has some very convincing arguments for why that is probably the case, and that even if it isn’t so, it’s worth a try.
3) US&UK have enormous soft power. Places they have never colonized sometimes have cultural cringe toward them. IIRC Alon said somewhere that this is Sweden’s reason for the experiment. Likewise, over the past two mayors, Budapest has explicitly copied TfL in some ways.
4) The purpose, the ends, of some part of the bureaucracy is itself a political football. For instance, the Green Party openly campaigns on opposition to highways, and whenever they win, they disassemble the DoT’s roads office. Their opponents, the strongly pro-highway Yellow Party, therefore implement their motorway construction program as a consultantified PPP to make it difficult for the Greens to undo it.
5) Some consultants already exist out there, and they will actively pitch you on building a monorail with consultantism (preferably, their company). There isn’t AFAIK any World Office of Bureaucracy who try to sell politicians the idea of building subways with in-house civil service.
If I’m reading the numbers correctly you are more likely to die of gun violence in the U.S. than you are to die in an … automotive.. “accident”.
The gun violence and car accident death rates matter, yeah, but the main reason American life expectancy lags is uneven access to health care in the working class. University graduates have about the same life expectancy at 25-35 in the US and France; for non-graduates, France is years ahead, due to universal health care. Bernie Sanders isn’t always wrong.
You didn’t say U.S. health care system sucks. Health insurance status doesn’t change whether or not you die by gun violence or automobile accident. Or opioid overdose.
you are more likely to die of gun violence in the U.S. than you are to die in an … automotive.. “accident”.
No, about 60% of gun deaths are suicide, and a few percent are accidents. Total gun murders in the us in 2024 were approx. 16,576, compared to approx. 38,800 in auto crashes.
Suicides went down in Britain when town gas which contained carbon monoxide was got rid of.
So if Americans had less guns there would be fewer suicides.
Perhaps, but not necessarily. Japan famously has both almost non-existent ownership of guns and a high suicide rate. The suicide rate in the US is high now, but close to Belgium, which is not a major gun owning country. Around 2000 US suicide rate was lower than almost every country in Europe, but it rose steadily from ~2005-2018 even though the rate of gun ownership remained steady or dropped slightly during this time (depending on source – note that number of guns in the US went up even as ownership went down since those who did own guns tended to own several instead of just one). In the 1990s US gun ownership rate was higher than in the 2000s-2010s but suicide rate was declining. There was a surge in gun purchases in 2020 with the pandemic, but suicide rate in 2022-2024 was only about as high as it was in 2018 beforehand (there was a slight decline in 2021-2022. In 2000 Switzerland literally gave each adult male a gun and ammunition to keep at home yet their suicide rate at the time was about the same as France and Austria.
There is a question why US suicide rate went up over the past quarter century while they went down almost everywhere else, but there doesn’t appear to be a correlation with rate of gun ownership.
Things which make it easier to commit suicide (e.g. guns) increase suicides.
The suicides aren’t dead? The overdoses are equally dead. I didn’t know there were gradations of deadness. The people who commit suicide by wrapping the car around a telephone pole don’t count? How about the people who deliberately overdose on the happy pills they got from the pharmacy?
They are all equally dead. And equally weighted in the life expectancy numbers because they are all dead.
They are all equally dead, but you said people were more likely to die of gun violence than car crashes. Suicides are not violence, just as overdoses are not medical malpractice.
Explain to us how getting your brains splattered all over is less violent when it’s a suicide than a homicide.
Because violence is physical force designed to hurt or intimidate another person. Suicide is done to oneself.
If you intentionally run someone over it is murder, if your tire blows out and you loose control and run someone over it is not murder. If the doctor injects 100mL of morphine instead of 10mL and you die that is medical malpractice, if you inject it yourself while trying to get high then it is not medical malpractice it is an overdose.
Events with the same mechanical or chemical action can be distinguished as different things, legally or practically.
If you want to parse things, “violent” is a different word than “violence” (violent being an adjective meaning destructive or damaging, or involving great force) so a suicide is not less violent than a homicide, even if it is not an act of violence that would show up in crime statistics anywhere in the world.
The statistics don’t care how pedantic you are. They are still dead. More of them by firearms than in automotive events. And more them by overdose than either.
The evidence base that Nordic costs have increased is pretty weak Alon, there isn’t much evidence on it in the transit costs project apart from maybe Oslo.
But then the cost difference there is similar to the difference between the two Helsinki projects that were done at the same time so 🤷♂️
Also I do think some care is needed with regards to the smaller projects.
If the preparation is weak e.g with the Battersea Power Station extension then some of the extremely high costs are about connecting that up to the existing line, and if it is designed for and partially built as can be the case then costs will be especially low.
Alon doesn’t just have Nordic costs, we deterioration across the Anglosphere (Canada), plus Paris adopting some of the globalised system. In the UK it happens in stages since its actually more important than the US in generating these practices. E.g. JLE doesn’t do design-build but the various DLR and Heathrow expansions do (I am currently trawling through these to update the database). JLE did have unprecedented use of outside consultants.
I think it does matter in terms of the recommendation we should give the powers that be.
If the recommendation is that Britain was fine pre 2010 then that is a different recommendation from Britain being fine pre-1997 and a different recommendation again from Britain only being fine pre-1992.
And there absolutely does need to be some account taken on project complexity which is presumably why the out of town Helsinki airport project is cheaper than the in town by the river Helsinki west metro project.
The Helsinki projects weren’t done at the same time – they’re two phases of an extension.
https://transitcosts.com/helsinki-ring-rail-line/ is 2009 to 2015
https://transitcosts.com/helsinki-west-metro-phase-1/ is 2009 to 2017
I call those two projects built at the same time.
Look at West Metro Phase 2.
How much did that second extension cost?
The only figure I can see is roughly €100m/km which is OK?
OK well I have looked at the source data and maybe you are right. But that said phase 2 includes an underground depot, so that is part of it.
Also it includes a bunch of expensive automation stuff.
Aren’t your norms of “professional civil service with no political overclass, strong political parties” contradictory? Wouldn’t a strong political party be a political overclass?
I know you have favored systems where the professional engineers plan the project and design its specifications, with the politicians only holding go/no-go authority, or deciding funding allocation thus determining how many projects can be built / how fast they proceed. Perhaps this is what you mean?
A strong political party is not at all an overclass! It’s strong in the sense of having a coherent set of longstanding institutions with a regular cursus honorum of elected politicians and an ability to enforce party lines top-down. It doesn’t require an overclass of political appointees. In fact the US has atypically weak parties – the DNC and RNC leaders are nobodies, party institutions are so informal that Trump could say with a straight face that Project 2025 wasn’t him but the unrelated Heritage Foundation (by German standards, Heritage is an in-party thinktank, like KAS, FES, etc.), senators break party lines all the time and until Pelosi came along so did House members, major legislation passes at the rate of about one per presidential term; it’s the strong-partisanship-and-weak-parties line used in comparative politics.
But yeah, what I mean is that in a system with strong parties (as opposed to personalism) and professional civil service, the politicians decide on overall funding allocation. “As conservatives, we’re going to cut the budget and say no to more transit projects” is what the right does in Sweden. They don’t delete the census website or try to cancel already-agreed individual projects.
Your disdain for democracy is showing. Again.
We tried that and decided we don’t like it. So there’s all the messy community outreach, scoping, environmental review and endless lawsuits. Much tidier to eliminate all the democracy and have a dictatorship of professionals. Assuming it’s the right kind of professionals because it could be the kind that want all roads all the time for all things.