Sunday, December 29, 2013

Predicting Weather, Complex Systems, and the Failure of Socialism's Centralized Planning


Predicting Weather
Predicting weather is tricky stuff. Weather patterns are extraordinarily complex and stochastic and never move exactly as expected. In today's world, we take it for granted that three day, and even five day, forecasts can come close to telling us what to expect. Beyond those few days, all bets are off. Even a few decades ago, predicting weather beyond three days was unheard of. One of the sayings we had when I was growing up was "I wish I could get paid for being as wrong as the weatherman."

Somehow we all know, in a common sense way, that predicting the weather, while scientific, is still inexact. We expect that predictions will have some variation and, occasionally, will be flat wrong. There's nothing wrong with the science. We understand that the weather system is too complex to predict much more than a percentage chance of certainty. (Hence, statements like "There's a 30% chance of rain today.")

Weather prediction has grown more accurate over the past few years because we have added more and more sensors across the US and the oceans. As more weather stations come online in order to measure local conditions, the network of a vast array of sensors can build a more accurate measurement of actual conditions. We measure wind, temperature, and atmospheric pressure at the ground level. Doppler radar scans for precipitation. Satellites give a bigger picture, measuring cloud cover, water vapor, and infrared variations in temperature. Since weather is such a complex system, we require thousands and tens of thousands of stations, gathering information, in order to accurately measure what the weather is doing here and now, let alone three days from now.

Complex Systems
It seems obvious to us that in order to measure such a complex system as weather, we must have information from thousands of sensors - the more the better and the wider the distribution the better. We would consider it folly to establish one, giant weather station in the center of the US (or worse yet, off to the side along the Atlantic coast, just north of Virginia) and expect anything like an accurate measurement of the weather. Understanding the system, let alone predicting it, requires thousands, tens of thousands, hundreds of thousands, of sensors to gather information and report the measurements.

And yet we understand that process is the only way we have in order to understand such a complex system, such as weather. Understanding other complex systems requires a vast measurement - and correct interpretation - of the data. For example, even after years of trying, no one really can predict the ups and downs of the stock market. Certainly, some are better than others at reading trends and then investing in trends to their advantage. Even more complex is the US economy, run by millions of businesses, manufacturers, services, all of which is handled by hundreds of millions of people. Our economy is a highly complex system.

Yet, somehow we lose our common sense when we talk about government and measuring (let alone predicting) the complex system of the US economy or the even more complex US society. It is a fatal conceit1 to assume that a single, centralized sensor has the smarts or the capacity to measure the myriad processes of a system as complex as society's. Let me spell it out for you here. Neither the president of the US, nor even the members of Congress can possibly measure and predict the economy and society. By extension, neither the president nor Congress can possibly direct those two complex systems without major errors.

Yet, somehow, we as Americans lose our common sense when it comes to handing control of our complex economic and social systems to the centralized planning of the federal government.

The analogy with predicting weather is accurate. Weather is best measured and predicted through a decentralized system. The economy best works the same way. In a free market system, businesses understand and respond to the local systems far better than any centralized agency ever can. A centralized agency can collect and correlate data, helping understand the larger economy as a whole, but should never assume that it can predict or manage the complexity of the system.

The Errors of Centralized Planning
We've known for years and have made endless jokes about the inefficiency of the bureaucracies developed by the federal government. These large bureaucracies fall short of expectations for the simple reason that centralized planning cannot measure, nor predict, nor control the complexities of a widespread system.

Take, for example, the smaller yet still complex problem of health care. In its fatal conceit, the federal government has assumed more and greater-reaching regulation of the local systems, assured in its vast hubris of being able to judge better than the free market, how to best run the health care system. The result is an even more complex system with greater costs, spiraling upward ever more out of control. It is not the "greed" of the market that has driven costs up but the regulation of centralized planning.

With the current health care laws, we can predict that the increased federal costs, bureaucracy, and regulation will damage the system. It may well completely break it. This must happen because centralization understands far less information about the complex system, not more. Lacking crucial information that it simply cannot obtain cripples centralized planning. The system operates nearly blindly, ignoring or stifling local information in favor of a large, "better" design.

A historical example will suffice to show the dangers we now face with the health care law. Lyndon Johnson created his own fatal conceit of the Great Society. He encouraged the US to create the welfare state and, among other things, federal regulation of local school systems. The intent of federal meddling with the local schools was meant to ensure a more equitable system for all Americans, black or white. The intent seems noble and good. Yet, after 45 years of federal interference in the complex system of education and after the "reform" of No Child Left Behind under George W. Bush, US students are doing worse in school than ever before. Illiteracy and dropout rates soar. Children are pushed through a system that resembles a factory more than a school. And now, with Common Core, we urge our educational system into greater and great centralization.

In a word, federal interference in all sectors of the school and education system has failed the education system in the US.

Modern leftist dogma continues to insist that the problem can be fixed as long as we pour more money into the federal system and regulate the system even more. The dogma insists that all will be made right as long as we tweak the system one more time. President Obama admitted as much in a speech at a Virginia middle school where he vowed to "fix" No Child Left Behind. As a sad symbol of this bankrupt philosophy, the New Jersey school named after Barack Obama is closed its doors.

The federal system has not, and cannot, fix the problem as long as central planning is the order of the day. The only reasonable way to measure and control complex systems is to decentralize planning, not to create an ever-growing leviathan of centralized government.

It's time to return to the doctrines of Adam Smith and the philosophy of free market. It's time to decentralize the federal system before it collapses under its own weight to take us down with it.

Notes:
1 The phrase is taken from the book by the same name written by F. A. Hayek, The Fatal Conceit: The Errors of Socialism, W. W. Bartley III, ed. (Chicago: The University of Chicago Press, 1989). Hayek presents a compelling argument against central planning and the welfare state, arguing instead the benefits of the decentralized system - namely, the free market.