Friday, February 24, 2012

Notes on Kling's PSST

Overall I don't think the Kling's theory (HT) is difficult enough to warrant all these words. Using that many suggests there's more to the idea than there actually is.
Being blunt, I praise it with faint damn. Yes, good work, you get to paste your name on this idea.

To confirm, I read it as: when technology shifts what jobs are profitable, it takes time for the people that make up the market to figure out the new stable equilibrium.[1]


In the real world, money is the only tool we have for measuring wealth. But when evaluating a theory, I can posit a non-recursive wealth measurement device, assume some concrete examples, and then see if the theory is correctly describing wealth dynamics.
"That is, unemployment can be blamed on an increase in the
demand for savings in the form of money, which does not require production."
I understand this argument as; people get thrown of out of work because others divert money from buying their stuff, so less stuff is demanded and fewer need to work to make it.
Loading up the wealth-analysis program, this becomes immediately, obviously false, because it is conflating wealth and the measurement of wealth. The logic is exactly as confused as saying the price of real estate went down because people wanted fewer metre sticks.

Verification one: imagine substituting my magic wealth-measure for money. Also suppose mouse-makers are losing work because people are saving rather than buying fancy new mice.
The mouse-makers want to sell mice so they can buy food. However, no physical input of either the mouse-maker or the buyer has changed. The raw resources, including labour, is all still there. There is no physical reason the mouse-maker cannot go on eating. Nor any reason the saver can't go on getting new mice.

Verification two: with money being stashed away, the same number of goods are demanding less money, leading to an increase in the price of money: deflation. The system's feedback automatically balances the supply of labour and the supply of goods, as long as the supply of the tracking instrument, money, doesn't change too rapidly.

In other words, the exact same stuff is demanded, but at a lower price. As long as the system has time to propagate the decrease, it won't be a problem.

Verification three: saving money is just a way of keeping score. It says; "Society owes me wealth in the future, because I forwent wealth in the present."
Sure, at first it looks like this will indeed reduce aggregate demand. However, it must be matched by a later increase in demand. More importantly, it is matching a previous savings, and increase in demand. For each yuppie who recognizes the value of a nest egg, there's a proportional old rural retiree who wants a new mouse. Again, as long as the shock isn't too large, the system automatically balances itself.

In short, no matter how I look at it, I can't find a stable mechanism for savings to be a bad thing. Turns out prudence is an unambiguous virtue. (Outrageous, I know.)

"Another approach is to suggest that wage rates can get stuck too high, leading firms to cut back on employment and output."
On the other hand, I have to agree with this. If wages get stuck, then you have a problem with prices and your economy will get sick, though I wouldn't be so bold as to so glibly assume the mechanics without first examining the feedbacks.

"when the rate of saving goes up, businesses do not invest to increase capacity in future output, but instead base their investment plans on what Keynes called “animal spirits.”"
[1]Now here's every reason I think I think Kling will have trouble coming up with further correct ideas.

This sophistry interests me, because it almost worked on me. It gives the impression that the whole paragraph is describing Keynesian thought, by the method of leaving all but one approach unattributed. However, I realized don't know for sure that the others are Keynesian, (though it seems likely) and Kling has plausible deniability in case they're not.

So...is it obvious that to use sophistry is self-destructive, even to imply true things? Though I don't blame Kling, sophistry is infectious and our culture is soaking in it.

"Each of these approaches to explaining aggregate excess supply is vulnerable on conceptual and empirical grounds. These issues of Keynesian macroeconomics were the topic of intense debate in the 1970s and 1980s."
It's for things like this that I perform my own analysis on the spot. Looks like Kling thinks I got the right answers: it is all Keynsianism, and it is all theoretically bunk. As a bonus, Kling believes the experiments back that up, and I'm willing to trust him on that.

Also it is a warm up for analyzing Kling's upcoming offering, and by explicitly working it out, it will help me recognize relevant details and distinctions

"Suppose that we had with us a time traveler from 1800.
Imagine taking a random sample of a dozen people working in different office
buildings and trying to explain to our time traveler how those people contribute
to the production process. Try to convey the role of a web programmer, a graphic
designer, a data analyst, or a social media marketing specialist."
Well, guys, I might secretly be from 1800. I'm not sure that those people aren't wasting resources.

For example, the designer makes prettier packages so that more of that stuff gets bought. Would voters have done without, if the packages were plain? Would they have been just as happy?


I am reminded that newer ideas aren't as well understood as older ones. It is likely that in the future, Kling will be able to usefully summarize PSST in one line just as he summarized various Keynesianisms.

"The PSST interpretation of the Great Depression in the United States would be that the internal combustion engine and the small electric motor disrupted patterns of specialization and trade."
If Kling were not a paid, 'respectable' economist, would he have included the dominating role of government in his analysis?
Does he omit it due to prudence, or due to actual lack of belief? In the latter case, if it were not in his professional interest, would he believe?

"The high school graduation rate was only 29 percent in 1931, but it had reached 59 percent by 1950."
I doubt I'll ever believe high school in fact causes skilled workers, for the unavoidable datum that I certainly didn't learn anything. The correlation is caused by the opposite causation. Future skilled workers more easily jump though the high school hoops.

I suppose a double-blind controlled trial - with replication, naturally - could change my mind. I expect to see the reverse result, though.

Come to think, this is like non-destructive quantum measurement. Because school doesn't do anything, academically speaking, you can be certain that it will never be tested. Just as you can be certain a puppy is there, as you never pull back a quantum steak.

The ones in the best position to perform a test are also in the best epistemic position without the test. If it would work, they would likely have already done it and crowed it from the rooftops.

On the contrary, what do you suppose would happen if you got to have (salad) lunch with the minister of education, and try to convince them to do controlled trials on education methods?

That said, it accomplishes Kling's goal upon his intended audience.

"The slump was a mistake, caused by bad luck and, above
all, by incompetent macroeconomic management."
Aware of government's possible role, but dismisses it entirely? That's...not good.

"On the other hand, could the patterns of specialization and trade in 1950 have been adopted sooner? [...] Even with such foresight, the patterns in 1950 depended on developments, such as road and highway construction as well as a more educated labour force, that were not in place in 1929."
Which means they can't have yet made previous patterns unsustainable, and thus can't be causing unemployment.


I suspect my most serious criticism is here:
"From a PSST perspective, one has to ask how full employment could have been maintained."
It doesn't need to be maintained.

It bears repeating: the ultimate goal is for everyone to be a Patrician and not have to work at all.

If fewer need to work to get the make the same stuff, unemployment isn't a problem, it's mission accomplished.


The solution is to substitute working time for leisure time. If unemployment goes from 0% to 20%, but the same stuff is made, the equilibrium is for everyone who wants to work to work 36 hours a week instead of 40, but get paid the same in terms of stuff.

There's an obvious bias that in principle should cause (all?) economists to neglect that leisure is superior to employment.
What actually happens when productivity per worker goes up 20%, is that the government raises taxes by a proportionate 20%, so that the workers still work 40 hours, make 20% more stuff but are paid the same in terms of stuff.

(Well, workers capture a fraction of the benefit - houses are bigger now, for example. Just not most of it.)

"I think that this would be a more damning criticism of PSST if the AS-AD approach were better established at predicting and controlling fluctuations in employment. However, the AS-AD approach does not produce reliable predictions of macroeconomic performance."
I agree. AS-AD is a lower bound on the usefulness of PSST.

However, the Austrians have an explanation. Unfortunately, the explanation fits neatly into Kling's governmental blind spot, and so he doesn't even see the need to dismiss it, let alone refute it.

It's doubly a shame because it also fits neatly with PSST. Subsidized debt leads to unsustainable patterns of trade. Which are then unsustained. Government doesn't learn, and returns to subsidizing debt. One business cycle for you, sir.

Again, using my magic wealth-measure: assume that prices in a freeish market economy do well enough at approximating wealth; then perturb the system by artificially lowering the price of cash; many ventures that are negative in terms of real stuff are now profitable in terms of dollars. This cannot possibly be sustained, as you run out of spare physical stuff to pour into the wealth furnace. The fire produces lots of GDP in the form of heat, but almost no light. Then the voters employed in shovelling coal for the train to nowhere get thrown out of work.

"Another difference was that the most recent recessions were followed by “jobless recoveries,” in which GDP returned to its previous peak long before the unemployment rate. [...] The “jobless recovery” represents a divergence between the path of output, as measured by GDP, and the path of employment."
I agree that this matches PSST theory.

It also matches my theory of government eating surpluses. GDP goes up, but voters cannot afford as much stuff, assuming employment is a good proxy for distribution of wealth. Where is all the stuff going?

This is where, if you're Kling, you should watch for confirmation bias. While the data matches PSST theory and therefore Kling can comfortably believe it if he so desires, a more diligent reading must ask; what else can explain this? Of these options, which interpretation puts my theory in the worst light? Does this worst case fit the data better than PSST?

Or, if you're me; is employment really a good proxy for wealth distribution? Where is all that stuff actually going? It may be going to voters. I'll have to check.

But the fact is I've repeatedly and antecedently noticed a legislative blind spot in Kling, one that moreover matches a blind spot I've repeatedly noticed in other professional economists. This is what I'd have predicted had I bothered to do so. I'm confident that he hasn't analyzed this possibility. Since I've analyzed both possibilities and he analyzed one, I'm in an epistemically superior position.

Still, this isn't an immutable advantage. Kling has shown willingness to take these things seriously when they're put in front on him.

"In the years leading up to the financial crisis, both the United States and the United Kingdom experienced expansions in finance, residential real estate, and government services. However, not all of the jobs were sustainable"
So close!

"The growth in finance reflected an excess of risk taking made profitable by misguided regulatory guidelines more than an increase in actual value added in that sector. [...] In this context, government spending is not likely to solve the problem."
But, definitely not actually getting there. Again, not a good sign. It seems Kling intuitively understands that government is a big issue, but won't consciously admit analysis of it.


Having written all that, I challenge myself to see if I really believe it. When I run across relevant data, does it make sense, or does it make me look the fool?


Update: (Pathway) Nick Rowe got the same conclusions on aggregate demand, though he did it the hard way. "since total output is the same, total income is the same. Depending on the elasticity of substitution, labour's share of that given income will either rise or fall, but absent any distribution effects, it's a wash at the macro level." The price of labour is determined by how much stuff there is to distribute; it's not a cause. Downstream, not upstream.
Note that not all of Rowe's points work when viewed as wealth mechanics.

Tuesday, February 21, 2012

Antibiotic Resistance, the Anarchist Solution

Coercive, government-legitimizing beliefs and thoughts are habit forming. It often stops the addict from seeing easy solutions.
"Tighter controls on prescriptions, or a tax on antibiotics, might address the conservation problem."
[...]
"If the market doesn’t work, why not try the government?

Even many libertarian types agree that the commons problem seems to call for stronger state controls over antibiotics." (Pathway.)
Principle: the market is smarter than you. Smarter than me, too, but in this case I apparently don't have to be that smart. Solving the conservation problem is straightforward.

Solution in 50 words or less:
First, make sure its health insurance, not healthcare insurance. The insurer should pay up whenever the customer gets sick. (How does this [ctrl-f "Stanford"] happen without collusion from the insurance companies?)

Next, sell the rights to antibiotics, wholesale, to the insurer. They decide who gets what chemical, when, and how much.

Verification:
The insurer has a a short-term financial incentive to use the antiobiotics, and long-term incentive to see that they still work later. The drug firms don't have to take on the market risk at all, but have an incentive to produce any chemical the insurers will buy. The insurers have an incentive to buy new chemicals that work, but only to use the cheapest that will work, and have the resources to professionally check that they work.

It aligns incentives for cooperation with the patient, all the way down the board. (Except the government, which sees an opportunity for overreach disappear.)


However, there is a flaw here; the third world.
"Laxminarayan likens antibiotics resistance to global warming: every country needs to solve its own problems and cooperate—but if it doesn’t, we all suffer. Coordinating a global response will require years, even decades; any serious revision to the patent system might have to go through the World Trade Organization."
They don't care about our insurance companies, and it's pretty easy to copy an existing drug. Resistance could develop there and then spread to here.
There's a straightforward non-combat solution, in globalizing the market better. This can't help in time, but is a nice illustration of how (principle:) past coercion destroys future non-coercive solutions, which implies that it is still worth ending current coercion to prevent this kind of thing from happening again.

Note, however, that all the solutions McArdle mentions have the exact same problem. (Are there any problems that markets have and governments don't? It seems unlikely in principle; they're both staffed by human beings.)


"The longer we ignore our problem, as Orwell did, the more likely we are to share his fate."
Who's 'we'? I'm on board with solving antibiotic resistance. By how much does that increase the odds that the problem will be solved, do you estimate?

Friday, February 10, 2012

Discovery and Verification of a Specific Epistemic Technique

I am discussing induction with Aretae and discovered - among other things - that my beliefs about the physical process of induction is consistent with my definition of intelligence.[1] It is a win-win-win situation. Also, two of the wins should be achievable through conscious effort, which means you can do this too, if I can figure out how I did it in the first place. It may even be automatic, which implies it is especially reliable.

Later this post switches to new discovery mode.[2]

I suppose I should think of a name for it, but it's a bit early as I don't know which one it will be.

If I'm wrong, it means my beliefs are being made consistent without conscious effort. It would mean they'll all shift if I find my mistake, without me having to spend any further effort. Verification: calculate the odds of two independent mistakes being made exactly the same.

If I'm right, either the beliefs are related or independent.

If they are dependent, it means I'm deriving new true things without conscious effort from previous facts. Verification: similar odds calculation as above.
If they are independent, it represents independent corroboration, and thus self-verification; based on the odds of a falsity appearing exactly consistent with a separate truth.


The next step is to get use out of this by figuring out which it is. If I'm wrong, in due course reality should kick me in the teeth either about induction or about my definition of intelligence.[1]

If I'm right, the possibilities can be distinguished by intentionally trying to use one of the resulting techniques. If I try to exploit automatic consistency generation, what happens? Do I get somewhere or do I get trash?

My present plan to to try these in order.

For completeness, I'll mention that the first time I try to exploit it for correcting false beliefs, I will do a manual re-check of all related beliefs. Problem is I'm running out of verifiably mistaken beliefs, as I've plucked all the low hanging fruit. There's the Higgs, (Via.) but I don't believe much that hinges on the Higgs...


[2]Wait. It's time to change this post to (approximate) stream of consciousness.

Thinking about the Higgs, I now have a new plan. I realized I'm wrong; I believe in the Higgs because I like the general relativity description of gravity as the curvature of space itself. You shouldn't need a particle to couple particles to space - it would mean the Higgs either was recursive or has no position.
If my beliefs are usually being logically coupled without conscious intervention, changing my mind about the Higgs should change my mind about GR without conscious thought. Historically, this has not been the case. Rather, I just get a flag next time I think about GR, reminding me about the Higgs. (And screwing up whatever I was about to say.)

Which means two independent beliefs are reinforcing each other. Neat. Verification: I'm waiting on the Higgs verification anyway; what they discover about the putative Higgs should be consistent with what they have previous said about Higgs. If not, it calls the entire model into question, and I go back to being happy with GR.

It doesn't even matter if my logic about GR is correct or not, it will verify the technique either way.

If Higgs is confirmed, I'll do what I have to accept it, then simply wait. How I feel about GR will either change on its own, or it won't - getting around the fact that I'll naturally remember to think about it, at least a little bit, now it has occurred to me.
If the Higgs turns out badly, I'll use the fallback plan.


[1] Intelligence, as commonly considered, is the conflation of three processes. Learning, creativity, and reasoning. This spans the space of everything you can do to bits; record them, generate them, and manipulate them. I believe they're implemented independently, at least in the brain.