"you going to set up an on-premises laboratory to test the quality of the concrete and steel?"
This isn't a good example because there are whole infrastructures to verify that buildings are being made "to code" and quality of materials is a part of that. E.g. this lab that tests steel: https://www.leica-microsystems.com/science-lab/steel-the-all...
Now what you're getting at is, OK, now you have to trust the lab. But that's OK because the incentives are better aligned. Trust is a thing with complex shapes. Maybe you develop trust in a single lab because being trusted is their whole business, and now you can easily switch between steel suppliers depending on who can sell the cheapest with peace of mind, so you've moved the trust problem around and reshaped it.
A lot of trust problems in the business world are like this. You can't eliminate it entirely everywhere, but reducing the amount you need and reshaping it/moving it around is still useful.
Yes, but even that is potentially troublesome. If you're in a low trust environment, then probably nobody trusts the labs either.
Like, why do the labs exist? Because the building companies really, really wants to build a solid building but the concrete industry is oddly shady? Or because there's an official/unofficial requirement for it? If the second, you easily end up with labs that rubber stamp whatever you want them to.
Trust in this sense is something that needs to be encouraged society-wide. You can patch imperfections, but there's a very real limit to how practical that is.
There's not really such a thing as low trust/high trust society. That's an abstraction that sounds good but hides too much detail to be useful when discussing specific projects.
You can't just encourage everyone to trust everyone else. To the extent developed countries can be described as "high trust", it's not because they are just magically more trusting. They get that way because there are extensive and mostly working mechanisms to find and penalize betrayals of trust. Corruption is low in the west because corrupt people get caught and punished, and because once the problem gets small enough people won't just blindly accept it but will make an effort to report it and ensure it's resolved (vs when it's endemic).
Stuff like concrete mixing is commodity, lots of small firms scattered around the world. They compete on price and delivery time so there's a temptation to cut corners. Delivery time is easily understood and measured, so is price, if you can properly measure quality then you can switch between providers easily. The alternative is that someone fixes the trust issue another way, like a big centralized mega-brand or having to in-source everything, but that comes with bigger downsides.
This comment helped me to put into words something I've been thinking about recently.
Propaganda that erodes the concept of truth in a society also destroys trust. While a dictator may benefit in the short term from confused opposition, their erosion of truth deeply damages the society on which their power depends.
That being said, I don't know how long a truth-optional trust-destroyed society can last. For instance North Korea, but I know nothing about how much corruption exists there.
This isn't a good example because there are whole infrastructures to verify that buildings are being made "to code" and quality of materials is a part of that. E.g. this lab that tests steel: https://www.leica-microsystems.com/science-lab/steel-the-all...
Now what you're getting at is, OK, now you have to trust the lab. But that's OK because the incentives are better aligned. Trust is a thing with complex shapes. Maybe you develop trust in a single lab because being trusted is their whole business, and now you can easily switch between steel suppliers depending on who can sell the cheapest with peace of mind, so you've moved the trust problem around and reshaped it.
A lot of trust problems in the business world are like this. You can't eliminate it entirely everywhere, but reducing the amount you need and reshaping it/moving it around is still useful.