I think that the general argument goes that something like "the economy" is something so immensely complex that any effort to manage it top-down down is likely to fail, regardless of how brilliant, talented, or well-meaning such efforts would be.
As analogy (imperfect though it may be), it's like trying to fix a bug in a program that consists of trillions of lines of code, except that you don't have access to the source code or a debugger. Instead, you have just a few high-level variables you can adjust (which may or may not be useful).
In summary, the issue is that there are some problems that have a scale and complexity for which a top-down approach is really beyond human comprehension.
Yep, I get the top down argument, but what i want to know is what does work, and why can't our government enable that?
Government is a massive bureaucracy, no doubt, but then so are large companies. And large companies can partner with smaller organizations, or buy them and use their knowledge and expertise to make money.
Government has the power of force, which includes the power to circumvent the price system by taxing or borrowing against future taxes, irrespective of the information the price system conveys about the efficiency of the actions undertaken with that money.
Since the government's actions may not be subject to the pressures of the price system, a government solution is less likely to reflect economic reality as described by the price system. In more than a few ways, this means the government's actions are not likely to succeed, because the primary motivations for experimentation won't be economic, but political.
Individuals and groups without the power of force must create solutions that can stand up to the pressures of the price system. In this model, shifting experimentation to the scope of individuals and groups by liberalizing (if necessary) or fine-tuning economic policy is what works.
Of course, this explanation requires an well-functioning price system, and making the system function well is a subject in itself. There's also the question of how often actors in the price system actually act rationally, given the information they receive from the price system.
As analogy (imperfect though it may be), it's like trying to fix a bug in a program that consists of trillions of lines of code, except that you don't have access to the source code or a debugger. Instead, you have just a few high-level variables you can adjust (which may or may not be useful).
In summary, the issue is that there are some problems that have a scale and complexity for which a top-down approach is really beyond human comprehension.