Let's say the Los Angeles city government has an empty lot they want to get rid of. They want something that will maximize economic return, in order to provide jobs and tax revenue (it might be worthwhile to dive into what measure of economic return they should use, but we'll ignore that for now.) Additionally, they're risk-averse and need to be able to generate public support for the plan in order to make it happen.

Now, which one of these plans will they choose?

• Engaging a large, well-known real estate developer who provides a nice 3D rendering of the shiny, multi-million dollar mixed retail/residential development he'll build
• Selling off subdivided lots piecemeal to smaller developers or individual owners

They'll always choose the first. The additional administrative complexity and risk involved in the second make it a non-starter. And, in this understandable way, they make the rich richer.

The government makes a ton of decisions beyond selling land that similarly impact the concentration of wealth, and an analogous process usually holds - if we need to regulate banks, for example, who should we turn to for advice? A million small players, or a few big bankers?

So, on average, government decisions will tend to be made in concert with large existing interests, and will tend to increase the concentration of wealth. I think this problem is structurally unfixable, and provides some of the justification for redistributing wealth.