For cheap infrastructure, we must lash ourselves to the mast

When it's not easy to do the right thing.

We all know how to lose weight. Yet few of us can do it (without the jabs). Five years after losing weight, studies say people regain an average of 80 per cent of what they lost.

We struggle to lose weight not because we don’t know what to do, but because we can’t stop ourselves from not doing it. One part of our brain wants to be skinny and beautiful. But the other part of it wants 9.5 per cent Belgian beers and sourdough pizza crust dipped in melted garlic butter.

Building a metro efficiently is a bit like losing weight. There’s not much disagreement about what should be done. But it’s hard to do in practice. One part of the government very much wants to do the right thing. But the right thing clashes with the priorities of other parts of the government.

Here are some of the common sense principles of efficient metro construction: make quick decisions; recruit competent people; don’t over-engineer; don’t commit to the project regardless of what it might cost; build a pipeline of projects so everyone is committed for the long run; don’t change the plan; don’t overpay when plans do change.

Metros go wrong because these principles clash with governments’ existing goals and principles. Here are some of them: don’t spend money before checking it’s a good idea; build the best possible thing; pay staff a fair, broadly accepted and pre-agreed rate; don’t commit to long term spending unless you have to; get reelected; don’t give up control; don’t make a mistake.

These principles aren’t wrong, per se. Like the module of our brain whose job it is to store as much fat as possible in advance of winter, they serve a purpose.

For example, Ireland has a department whose job it is to stop the government from wasting money. It carries out this task by monitoring and testing spending decisions. The problem is the principle of monitoring and testing spending decisions clashes, in the metro domain, with the principle of making quick decisions. When something unexpected happens that necessitates the spending of more money (which happens a lot when you’re tunnelling under a city), a quick decision is much cheaper than a slow one. Testing a decision at length delays the answer and adds cost. The act of trying to control spending ends up causing wasteful spending.

There’s another way the impulse to control spending adds to cost. It’s to do with the duration of funding. If the project needs to get its funding renewed annually, that gives the finance minister and his officials leverage over the project. They can demand changes every year at budget time. This conflicts with the principle of well-run metro projects that plans, once made, should not change. Another problem is that annual spending derogations take leverage away from the project in its negotiations with contractors.

The arm of the government that manages the purse strings will be inclined to under-invest. Other arms of the government, by contrast, are unconcerned with money. Their job is to build things. Their inclination is to build the thing as well as possible, at any cost. They are inclined to over-invest. This tendency, too, is in tension with cost-effective metro delivery. Italian, Danish and Spanish metro stations tend to be small, plain and functional, in contrast with cathedral-like stations found in English-speaking countries. A well-run delivery agency shouldn’t spend more than necessary to deliver the function. Gold-plating is the enemy.

Another conflict is between the civil service’s (rational) impulse not to screw up decisions and metro projects’ need for quick decisions. A perfect decision – in the sense of the one with the lowest error – is, in practice, a bad one if it delays the project.

Another conflict is between the government’s impulse to minimise long term spending commitments and a metro project’s need for qualified staff. The metro needs specialists whose first loyalty is to the project. Governments are often tempted to hire consultants on a short term basis, and in doing so avoid the long term commitments of full time employees, salaries and pensions. But the reliance on consultants, as opposed to in-house staff, is associated with higher project costs.

A related conflict is between governments’ impulse to stick to settled pay scales, and metro projects’ need for specialised talent. A metro line might cost €15 billion. So the difference between good and average leadership could be worth billions. As a result, experienced project leaders command a big salary. Governments, on the other hand, employ hundreds of thousands of people, and understandably are not inclined to negotiate salaries on a case by case basis.

Another source of tension is between ministers and metro projects. Ministers are democratically elected, and are the ultimate clients. They have control over the project. But ministers’ want to get reelected. They will be tempted to tweak and adapt the project to help advance their own aims. This obviously causes problems for complex projects, which work best when they stick to a single plan.

In the UK, the HS2 project has suffered from continually shifting direction from cabinet. The Stewart Review into the governance of major projects said it: “has been subject to evolving political aims, which pushed forward on the schedule before there was sufficient design maturity and caused progressive removals of scope.”

“Another consequence of the politics has been a difficulty in making decisions, leading to major delays and additional cost,” the report added.

Lash ourselves to the mast

Conway’s law says “Organizations which design systems are constrained to produce designs which are copies of the communication structures of these organizations.” A pithier way of saying this is that organisations “ship their org charts”. What this means is that companies’ internal politics — rather than the needs of end users — determine what gets built. This is as true for governments as software companies.

The bad news is that it’s hard to go against this tendency. Like the dieter struggling to resist strong beer and oily bread, it takes an exceptionally well-run project to go against the organisation’s nature.

The good news is that it’s possible to create new org charts and institutions that are compatible with the needs of complex projects. Like a jab of ozempic, these governance setups block unhelpful impulses before they cause problems.

The basic idea here is pre-commitment. Before a government decides, for example, what kind of concrete to use, it should decide on the right processes for making these decisions. It should pre-commit to good decision-making frameworks.

The government should pre-commit to delegate spending authority to the project. When a decision needs to be made on the fly over what type of signalling equipment to install, it should be made quickly by qualified and accountable experts in the project leadership. It shouldn’t be decided slowly by risk-averse and non-expert officials.

The government should pre-commit to a pipeline of future projects. It’s harder to get value for money on a once-off project than an equivalent project that’s part of a bigger pipeline. The promise of the pipeline incentivises more contractors to join the market, and justifies the state’s investment in its own talent.

The government should pre-commit to itemised contracts. This is a way of controlling costs when plans change on the hoof – as they inevitably will. It’s quicker, easier and cheaper to pre-agree the cost of concrete, labour etc, than to negotiate with contractors every time there’s a change.

The government should pre-commit to expertise. It should hire and promote skilled delivery experts. Some of these will be brought in on short term contracts. Others will sign on for the long term with all the associated pensions liabilities and salaries. This cadre of experts should be tasked with delivering the project.

The government should pre-commit funding, ideally for the whole duration of the project. This is how Copenhagen’s metro was funded. The fewer times the project needs to return for fresh funding the better. Every round of fresh funding is an opportunity for the rest of the system to meddle with the project.

The city of Madrid built its metro exceptionally quickly, cheaply and extensively. It followed all the common sense advice set out above: it didn’t over-engineer, it built quickly, it didn’t change plans once it had started.

It’s probably not a coincidence that Madrid is endowed with a government org chart that’s well set up to avoid the traps I described above. The city has spending power, power over planning and permits, and it has a mayor who’s incentivised to get the project built at a fair price. To extend its metro, the city of Madrid spun up a special purpose vehicle called MINTRA, at arm’s length from the rest of the city government, to help resolve these tensions.

When it comes to building complex projects in Ireland, delegating political control to Madrid-like city states is probably a bridge too far. But if we’re aware of traps in advance, and position ourselves to avoid them with SPVs and other pre-commitments, we could get a similar result.