One very Hayekian (Austrian) reason is that you won't have [very useful] data on consumer preferences. The price system will aggregate this effectively. You can't really poll for this, but purchase decisions will reveal it.
UPDATE: And this applies not just at the consumer level, but in the decisions firms make regarding their supply chains.
In what way do prices help with that? If more Fuji apples are requested, you know people like that over the other varieties. If you want one brand of apples over the other, one brand's products_distributed variable increases by one. Sorry if it appears as though I'm being obtuse.
>And this applies not just at the consumer level, but in the decisions firms make regarding their supply chains.
Howso?
Maybe I prefer Fuji apples, but only slightly; I'd be perfectly okay with a Nashi pear too, or if there's a shortage of Fuji apples, a Royal Gala.
Maybe a frost has reduced Fuji apple production; it would make sense for Fuji apples to be allocated first to the people who really prefer them, and then if any are left over they could come to people like me.
> If you want one brand of apples over the other, one brand's products_distributed variable increases by one.
Right, so how do you capture my "sure, a Fuji apple, but only if there's plenty"?
In a price system, it's easy: The price of Fuji apples fluctuates based on supply and demand, and I'll purchase the fruit based both on my internal, ever-changing preferences and the price. Regardless of the changing relative scarcity of Fuji apples, I will, apparently magically, end up with the fruit that more efficiently matches my wants and needs with the global fruit supply.
And this works even if you scale it up to 7 billion people and millions of products. That's a very difficult trick to match!
(And keep in mind, we also need to make decisions about distribution, stock levels, what apple types to grow, how much land to allocate to orchards, how much fertilizer to use, where to locate orchards to maximise growing season but minimise fuel usage for transportation. Every resource we have is limited in some fashion, often sharply so, yet our wants and needs are effectively boundless.)
That said, planning on top-level is still needed and still done. The point of setting up feedback loops isn't to let them run unchecked, but to make it easier to control the whole system (including keeping it stabilized). Just like it would be stupid to build a thermostat without input, having the temperature be determined by the point at which the materials it's made of start to "sweat", you don't want to leave market totally unchecked, doing whatever the hell it wants. Governments still get to influence the market, the same way a temperature dial influences the thermostat.
--
[0] - as performed by humans with pen and paper; I don't see a reason why a centralized planning scheme could not work better than capitalism if performed by optimization algorithms in a hyper-connected world.