More centralised authority, no local counter initiatives that bog them down, less political gridlock and stronger consensus mechanisms, less influence by third parties through lobbying which the US glamorises as free speech, no bias against public transportation and so on.

The US almost prides itself in the fact that its government is dysfunctional, that individuals can stop projects at the expense of the general public. One just needs to look at the countless of Hollywood movies where some lone Erin Brockovich-esque hero stands up to 'the machine', with the question rarely being raised what the collective costs and trade-offs are, or whether there's even a rational case to be made for the resistance.

There is no big problem with going through courts to make sure the project has sufficiently good cost-benefit ratios.

The problem is, that alternatives are not priced comparably, thus the economics simply favors them. (Such as driving or flying.)

If tunneling would be cheaper (in concrete construction costs), then those could go ahead, because there's not much issue about it, it's about as straightforward as flying. (Of course below a certain depth, so no noise and structural concerns arise.)

The people managing these projects are often hired based on political connections, not expertise. And cost overruns usually lead to higher profits, so there is little incentive to keep costs down.

While I’m always a fan of remembering that incentives matter (and believe that some degree citizen attention is a price for a well-functioning government), I would think that a forum full of software engineers could think of other reasons projects go over budget than graft.

It's not graft exactly, but simply the project structure leads to enormous inefficiencies. Eg subcontractors cannot work/meet/organize/plan directly with each other, because financial interests put silly lines in the sand. (Just think about how slow/dumb things can become if the backend, frontend and ops are completely separated, plus there's a middle layer (the general contractor) between them and the stakeholders.)

Plus, it seems that the main problem is that there's simply not enough experience/expertise at governments for managing/planning/executing these projects. Because too few people are in-house (or on payroll/retainer with the right/aligned incentives), and too few projects are done, so no competent team forms over time.

> I would think that a forum full of software engineers could think of other reasons projects go over budget than graft.

We're not talking about projects in the private domain. We're talking about government projects. Look at the Obamacare website boondoggle. Or the recent lack of backups in Baltimore when hit with ransomware. When it's a government project, it's very fair to consider graft and corruption as the primary causes, especially as they often lead to favoring incompetent people in charge of executing the project.

Is this a problem unique to the US? Or is it just worse in the US than in Taiwan?

Sort of. The US is big and rich (so small mistakes compound into really big expensive errors fast/eventually) and sort of refuses to shop abroad for expertise, because. Reasons.

I'd say in the US at least it's hard for a politician to devote money toward a multi-year program that their constituents won't see the benefit of for quite some time (or in the case of upkeep and maintenance, at all).

The politician isn't looking to do what's right, they're looking to get reelected.

Why are politicians in other countries able to accomplish this while those in the US can't? Are American politicians just lower quality than those in other countries? Or is that their constituents are more prone to short-term thinking than their counterparts elsewhere?

> The politician isn't looking to do what's right, they're looking to get reelected.

But that's true in all democracies, pretty much by definition.