Wednesday, September 17, 2014

Keynes, Einstein and Maxwell

People like using "wise quotes". Certain phrases are repeated continuously, each time with the user expressing a self satisfied look, like he'd really nailed his argument into place, and was oh so intelligent.

Often, the actual people who said the line are far more intelligent than the people who repeat it. As such, the lines, used as the author did, are often much more sensible than their common usage.

In this case, I'm not so much concerned with arguing against the (original intent) so much as the (mainline message).

That said, I'll introduce our wonderful Keynes quote:
"In the long run we are all dead"

Time is dimensional. This is required to make basic tenets of morality work. Consider murder- A man aims and shoots a gun at a random person in a crowd with no justification.

How does this work?

There's no frame of time in which he killed his victim. Also, there is a gap between the frames where he was applying force to the trigger, when the bullet left the gun, and when the bullet hit its target.

The common idea that time dilates morality, would mean that, between two shooters, the one who shot at people further away from him, is better than the one that shoots at people closer to him. It would mean that someone who salts anti-personnel mines in a forest is better than someone who opens fire on a crowd.

Furthermore, if time dilates, then distance should also dilate, because the two concepts are congruent. As such, a person who shoots his neighbors is worse than someone who fires some sort of awesome laser weapon into a crowd of aliens in another galaxy.

If the universe is going to end from entropy some day and a new universe is going to be born in the distant future, and some mechanism can be used to predict and influence the new universe from our current universe, and if that new universe is going to give birth to an Earth like planet, it is unimaginable that hurting people on that "new Earth" would be somehow more acceptable than doing the same on "current Earth".

Morality should not contain arbitrary delineations. As such, it is absurd to think that morality would dilate over time.

With this said, when would morality, or at least the calculations that are used to apply morality, dilate?

The answer is- data and processing power.

Now we've reached a standard that makes sense. Between shooting someone right in front of you, and shooting bullets into a thicket, obviously the thicket is the more non-violent choice. All actions must be taken under a backdrop of information scarcity, and so, utility calculations must account for "fog of war".

Processing power is similar though different. Even if we have sufficient data, that doesn't make it usable, or worth using. Sometimes, we can't reach the correct answer. Other times, we could, but it isn't worth it. Everyone should take a look at Bayesian probability, and incorporate its key concepts into their systems. However, actually drawing up a Bayesian tree should only happen in the most extreme cases.

Time and information tend to be directly correlated.

However, this is not (fundamental law) so much as a (common flow).

There are times when we have more information about the far future than the near future. For instance, imagine a game where red or blue marbles are randomly added to a pool each turn. On turn three, what is the standard deviation of the ratio of blue to red marbles?

In turn, consider the standard deviation on turn 1000.

This phenomena is frequent. The close and the far become "reality" whilst "darkness" lies in between.

To dismiss the far, and concern ourselves with the between instead, is absurd.

At times, far data is as reliable, or sufficiently reliable, that trade offs can be made between far and near. For instance, it's a decent bet that if you keep a stock of gold bars around, it'll benefit you someday. If you instead take the money and eat some restaurant food, that will also benefit you. Sometimes, gold markets crash. But, sometimes restaurants make their customers sick. There is no choice that entirely avoids risk. The data has to be weighed, and prioritized.

Of course, economics is the same. Expanding the money supply increases market noise. Government spending increases market noise. Taxes also increase market noise. Note that not all taxes, or spending programs, or currency issuances,  are the same. Issuing 5 trillion dollars into Ron Paul's bank account would probably result in less market noise than issuing a million into Paul Krugman's. A citizen's dividend (with equal funding) would cause less noise than Social Security. Every time the tax code is made more complicated with weird rules, exceptions and social engineering schemes it generates more noise. A simple tax that took 90% of all owned currency every year, would, assuming none of the proceeds were spent, generate less noise than our current tax code. (basically, a system wherein every year, dollars cease to be called dollars and are instead called dimes, pennies tenth pennies, etc.)

Because government is in constant flux, and because it's hard to tell how a program will operate after implementation, it's difficult to determine how much noise a given program will produce.

However, since noise is a fundamental element, work cannot be achieved without it.

This leads into a severe flaw of mainstream economics-

Consider Maxwell's demon.

The thought experiment began as an attack on the second law of thermodynamics, and it was one of the most interesting lines of attack. After significant effort (continuing into the modern day) scientists finally proved why, Maxwell's demon cannot exist within Newtonian physics.

The important element of their solution was that, for information to exist within our world, it required physical backing- aka, acquiring the vectors of a molecule requires payment, both in energy and entropy.

Under these principles, we can build a "demonic engine" that builds and releases pressure from a chamber, by opening and closing a door, and thus produces more energy than is used to open and close the door. This is made possible, because information can be transmuted into and from energy.

Mainstream theories consistently try to prove things in overly abstract "clean rooms". For the most part, I do not disagree with the idea that spending more during a recession will reduce the severity of the recession, or that moderate inflation will encourage more work to be done, etc.

The problem here, is that the economists are cheating. They are becoming Maxwell's demon. It is easy to come up with a program that improves the economy when you receive free information which you can then feed into the economy.

Without any information input a contracyclic spending program will still reduce the frequency of recessions. This is because-

If you flip one coin, the chance that you will get a tails is 50%

If you flip two coins, that chance of two tails is 25%

75% of the time you will not get two tails.

50% of the time you will get 1 heads and one tails. IE, either the government will be stimulating the economy, or the economy will be booming.

On the face of it, stimulus will avoid recessions.

However, a particularly disastrous scenario is prepared-

Stimulating the boom instead of the bust.

The problem is, this is just shuffling costs into a larger pool. As a result, a few long drawn out depressions replace many six month "soft periods".

Now, most economists would assert that working without any information at all is unrealistic.

I disagree, on a simple basis-

If any investor bought at the start of booms, and the sold when busts began, he'd be far richer than Warren Buffet. If any investor bought whenever a boom was going to continue, and sold whenever a bust was going to continue, he'd be similarly rich. Even now, economists argue whether we're in a boom or bust. It is obvious, that, determining whether the market is currently in a boom or a bust is impossible.

Even if someone did have the skill set necessary, that still does not mean he would be working for the federal reserve. Or that he would be a professor from Harvard. Or etc. Before the government can state that it knows something, it would have to have some method by which it can identify people with special skills.

I agree that, economists do not have to be speculators to be taken seriously. Mises never attempted to make himself rich. Nor did Adam Smith.

However, economists must be careful that they do not attribute to themselves (or worse, the general public) skills that would make them top class speculators.

In other words, much of what Keynesians use to "prove" that government can make us richer by managing our wealth, merely proves that god almighty could make us richer by managing our wealth. Not that an arbitrary or democratic bureaucracy can make us richer.

Which in turn asserts a particular axiom-

No government program will ever be implemented in the manner described by theoreticians.

This axiom is consistently true because, the theories it is applied to exclude three critical factors-
The incompetency of bureaucrats, the stupidity of crowds and the incredible complexity of problems existing in the real world.