Walter Williams' most recent column, while apparently stating the obvious, is still good to read. Why? Because people still seem to believe that we can manage the economy in some sort of scientific way. As I wrote in my post "The Limits of Technocracy", long ago, it is simply beyond our ability to understand, much less manage all the interactions in the economy. And that is precisely the point Dr. Williams makes.

Of course, some people, those who have little interaction with computers, have been bamboozled into thinking technology now allows us to understand things many orders of magnitude more complex than even our best computers can handle, be it the weather or the economy. Of course, even if our computers were many times more powerful, we still could not truly predict the weather or manage the economy.

Why not?

In the case of the weather, the answer is simple. While the weather may obey mechanistic principles, there are simply far too many parts interacting for us to easily model it. Any model is of necessity a simplification, and loses some predictive value. And that is not all. Even if we had the computing power, we simply do not know all the forces yet. Every model makes assumptions about what does and does not contribute to the weather, but those assumptions are nowhere near certain. And so our computer models, even assuming much more powerful computers, are inaccurate simply due to our lack of knowledge. We just don't know everything that causes weather cycles, so we cannot accurately model it.

The economy, on the other hand, is an even worse candidate for modeling. As I wrote in "The Limits of Econometrics", the economy is not mechanistic, it is based on the subjective valuations of every individual. Not only are those capricious and subject to change every second, they are also not numeric. You do not want carrots at level 4 and potatoes at level 6, you just want potatoes more than carrots. Maybe a lot more, maybe a little more, but nothing to which we can attach numbers. And that means that any attempt to model the real source of valuations is impossible, as dealing with ordinal rather than cardinal numbers is beyond the ability of computers.

Even if we abstract somewhat, and instead use supply and demand curves, ignoring individual preference by asking "How many carrots would you buy at each price", it does not solve our problem. The problem is precisely the one Dr. Williams describes in another form. Each curve is valid only as long as every other curve remains stable. If potatoes go down one cent, it adjusts every individual valuation of every other product. And that in turn changes the price of every other good, which necessitates an infinite adjustment of every curve for each change in the other prices. In other words, we have billions or trillions of curves for each individual, with a change in one requiring changes in the rest. However, it doesn't end there. Since the market price is determined by the combination of all individual curves, these trillions of adjustments in individual curves will change the overall price. And when price changes that once again changes individual preferences, as relative price may change one's opinion of the attractiveness of a given good.

In short, those nice clean supply and demand curves from ECON 101 were lies, it is a messy web made of millions of individual messy webs. We can approximate it with cleaner diagrams, but those are abstractions, and being so far abstracted from the truth are quite far from accurate as well.

And even if we could manage to solve this problem, we encounter the final, and probably insoluble problem. Even if we could model the incredibly complex interaction of all these valuations, it would be meaningless, as relative valuation, the subjective value which determines all these curves, can change every second. And so we would have to repeat our "how much at each price" survey for every good for every consumer each and every second in order to keep abreast of changes in valuation, as, otherwise, our graphs would become increasingly inaccurate.

And that, truly is the problem with anyone who assumes they can mange the economy better than the "chaotic" free market. Yes, the market does not perfectly respond to every change, but it does allow each individual to pursue their own preferences in the manner they see fit, which maximizes their own satisfaction. Any "scientific" management, claiming greater knowledge of what should be done, is destined to fail, as it is simply impossible to know enough to make all the adjustments needed. At best, the central manager would reproduce the slightly lagging free market response and thus be useless, far more likely it would produce worse results and leave everyone less happy.