« Reading List: Agent Zigzag | Main | Snowman »

Saturday, November 22, 2008

Gnome-o-gram: “Managing the Economy”

One incessantly hears the legacy media prattle about the U.S. President or the Federal Government or the Federal Reserve or some other instrument of coercion "managing the economy". Have any of the people who write or publish such nonsense, or those to whom they impute such powers ever managed anything at all of substantial size? Dunno—but I have. I was founder and CEO of a company which went public in 1985 and today has a market capitalisation greater than that of General Motors. This is not to brag, just to cite experience relevant to the observations which follow.

Let me tell you what it's like to “manage” an enterprise vastly smaller and far more easily directed than these railroad-era continental-scale debt-financed fiat money economies the politicians pretend to steer. The legacy media and, I suspect, politicians who haven't been there, assume the economy is something like a ship. On the bridge is a steady master, with a crew responding immediately to helm and engine orders, with progress plotted continuously on accurate charts.

In reality, it's like this. You, Mr. or Ms. CEO, are sitting in your office. Your desk has dozens of levers you can adjust and dials you can twiddle affecting the disposition of financial resources within your organisation. Half of these do nothing; a third of the remaining have results opposite to your expectations; and the balance work in the expected direction, but with disparate and often nonlinear effect. All of these controls, for better or for worse, have no immediate effects upon visible results, but only after a lag which is often unknowable and interacts with the settings of the other controls. And you have no idea which of the controls have what kind of effect upon the results.

Your information about the current state of affairs and the effects of your adjustments to the controls comes as measurements of financial aggregates which necessarily discard much of the detail subsumed into them. These aggregates are reported weeks or months after your changes to the controls are made, and you can see only the net effect of all your changes, not those due to each individual control, nor the lag times between adjustments and results.

You know next to nothing about the inputs; you have a large number of controls about whose effects you are ignorant; and your feedback from results is ambiguous, impossible to trace back to the individual actions which caused them, and delayed by an unknowable interval from cause to effect.

And you're supposed to manage that, and be responsible to the shareholders for the results? Now you may begin to grasp why CEOs get paid so much. Heck, when I was CEO of Autodesk, I was paid sixty thousand dollars a year to purport to do that, and, if I say so myself, I earned every nickel of it—well, most of 'em, anyway.

Now consider how this scales. When you're running a small business, as I did from 1978 through 1982, you can keep everything in your head: cost of goods, inventory, key customers, quality issues, and development priorities. This is one of the main reasons small businesses are so efficient and form the backbone of every genuinely free economy. When you get bigger, there's so much going on it doesn't fit into the head of a single individual any more: you need to develop management tools to abstract the essentials from the raw data and highlight the priorities from which strategic decisions are made. As you get larger and larger, the “economy of scale” argument is that this is logarithmic and hence as the enterprise grows, the relative overhead due to management and administration decreases, but I would argue that it becomes geometrically more difficult: you are trying to aggregate more and more disparate data into each single item on a financial report, to be evaluated by somebody who has little or no knowledge of the detailed components which make up that number and yet must make decisions based upon it alone.

Keep in mind the geometric growth in distance between reality and available management information as the size of an organisation grows. When you scale this up not just from the operator of a family business to a mid-sized publicly-traded company, but all the way to a national economy, the exponent kicks in hard, and it kills. The would-be “managers of the economy” stand before a vast control panel with thousands of unlabeled knobs, none of whose effects are known, bedizened with a multitude of gauges indicating nobody-knows-what-precisely, with unknown and unknowable delays between whatever they measure and what they display.

Okay, go manage that.

Which is why whenever you hear folks talk about “managing the economy” you should chuckle and place your assets outside of their reach and as uncoupled as possible from the consequences of their actions.

Other gnome-o-grams

Posted at November 22, 2008 21:42