This is the second part of a review of George Gilder’s monograph The 21st Century Case for Gold: A New Information Theory of Money. The first part, about the return of a gold standard for money, is at The Stream. You need to read that part before continuing. We return to SAMT in two weeks.
Gilder doesn’t have a discussion about how a gold standard can be implemented. He steers mostly clear of politics and sticks with theory. His monograph’s subtitle is “A New Information Theory of Money”, but perhaps it’s better classed as an intriguing new notion. In it’s current form, the theory is a thin skeleton that badly needs flesh. And a few corrections.
What’s right is his search for a definition of what money is. Time, he says. Time is “irreversible, inexorably scarce, impossible to hoard or steal, distributed with remorseless equality to rich and poor alike. As an index of time, gold imparts the accurate price signals needed for sustained economic growth and expanded opportunity.”
This makes sense when it is considered early humans “had every natural resource we have”, and the only difference between them and us is knowledge, or information purchased by time. While this is true, or true enough, it’s not clear what function of time equals wealth. Whatever this function is, it almost certainly can’t be linear; there must be some threshold effect; wealth must “kick in” only after some bare minimum of human needs are met. And that minimum, given our biology, must be food. It’s only the well-fed who ponder wealth.
Gilder is keen to map time to wealth and wealth to knowledge, or information. It is accepted by everybody that information is a key to the economy. Ideas drive the economy forward—and backward. Information is a value-neutral term. The government, for instance, is continuously presenting the economy with new “information” in the form of regulations, laws, mandates, and so forth.
Gilder’s enthusiasm for this fecund line of thought leads him astray. He says things like Claude Shannon, a prime inventor of information theory, “resolved that all information is basically surprise. Unless messages are unexpected, they do not convey new information.” And “Surprisal—what Shannon called ‘entropy’—is both a measure of freedom and criterion of creativity.”
This is muddled. Shannon did not define information as surprise, a claim Gilder makes several times. Shannon was interested in how information is communicated, not in information per se. Information transmitted via some channel is, by Shannon’s rule, “surprising” to the extent it is improbable. But to be improbable means the fundament of the information must already be known, and that is because all probability is conditional on what is known or assumed.
Consider transmitting English letters (in sentences, say). The letter z is rarer than e, so it is more “surprising” to see a z. The probability of seeing a z, and all the other letters, has a bearing on how to best encode the information so that it arrives at its destination in the most efficient way possible. And we know the alphabet in advance. You cannot transmit an unknown letter, without also including its definition.
To some extent, information transfer is how the economy works, which is plain. But surprise isn’t everything. Information that is known can be more valuable than new information which is surprising. After all, it is already-known information that lets factories produce their goods. Information theory does not, and can not, account for the creation of information, which is by definition infinitely surprising. Much more work needs to be done in figuring how to best encourage the creation of new, good, and true information.
Gilder also says the “second law of thermodynamics ordains that entropy as disorder always increases and cannot be reversed” and that this, somehow, ties to the “time-based entropy of [gold] extraction.” Now it’s true that entropy for closed systems, like the universe as a whole, increases through time, but it is false that everywhere entropy increases. Every new good idea produces new information, as Gilder rightly emphasizes, acts which decrease entropy, or rather, increase order. Entropy in the information-theory sense increases when unpredictability increases. The economy is not a closed, or zero-sum system.
The difficulties multiply because we can’t say we’d unconditionally like to have an predictable economy. A Soviet five-year plan was highly predictable, but produced an economy only New York Times readers would enjoy. So we’re not chasing entropy, high or low, but good information, knowledge. Knowledge is information which is true.