Less Wrong is a community blog devoted to refining the art of human rationality. Please visit our About page for more information.

Comment author: JoshuaZ 23 November 2012 04:18:07PM *  1 point [-]

Inferring causality from a time-series of various economic variables is incoherent. There first needs to be a deductive understanding of what the causal relationships are between economic variables. That is what is meant by "first principles" here - perhaps the disagreement is semantic.

If I replace "economic variables" with "astronomy" what part of this sentence changes in implication? Why is this incoherent for some fields? The level of rigor cannot be the only difference: Physics and astronomy have become more rigorous over time, not by discussing first principles, even when moving from Aristotle to Medieval physics to Newtonian physics, but rather by adopting principles based on the empirical data. The same goes for the switch from Ptolemaic systems to Copernicus and then Kepler. It is the empirical data that matters, the apparent time-series of planetary and stellar motion.

Thus, the first principles analysis: What is an economy?

Does arguing over this definition pay any rent?

Comment author: Kal 26 November 2012 02:23:20PM -2 points [-]

Re Physics, please correct me as required but in the way I use the phrase "first principles" here, Physics does not have any first principles. Physics is observation, hypothesis, experimentation and repeat. After a certain hypothesis has sufficient amount of experimental proof behind it, it becomes a theory and thus the foundation for further work. And occasionally, we find that there is a variable missing in the theory as the experiments did not test the situations that that variable speaks to. Then we test to tease out the nuances of that aspect of reality. And so on.

Economics has first principles, in the sense I use the phrase. Thus the Q: What is an economy? It leads to those first principles and then deduction covers the rest. But one can of course get the first principles wrong and the deduction is then useless.

Comment author: pragmatist 21 November 2012 09:20:47PM *  12 points [-]

The leap from "controlled experiments are not possible" to "one has to deduce from first principles" is huge and unsupported. The results of controlled experiments do not exhaust the available empirical evidence by a long shot. We have a lot of data about the effects of monetary policy from around the world. True, inferring causality from this data is not nearly as straightforward as inferring causality from a randomized controlled trial, but it's still a lot more reliable than deduction from first principles, I would think.

Think about how your argument sounds when applied to cosmological theories about the very early universe. We have a number of different theories that cannot be simultaneously right, and we cannot conduct controlled experiments. Would you endorse deduction from first principles in this instance as well?

Comment author: Kal 26 November 2012 01:59:26PM *  1 point [-]

...it's still a lot more reliable than deduction from first principles...

I know it's been a while since this comment, but I wanted to comment that sound deduction (ie, logical reasoning) on top of correct premises trumps other forms of evidence, even well-controlled, large-sample experiments. Or put another way, the former renders the latter superfluous. We should not simply take anyone's word for the deduction of course - we have to double-check the premises and understand the deduction step by excruciating step. But once one is satisfied that the premises are correct and the deduction is bullet-proof, one can sleep soundly :)

Nothing trumps sound reasoning. The scientific method itself is a result of reasoning, a point well made by Moldbug here: http://unqualified-reservations.blogspot.sg/2010/06/three-homeworks-for-professor-hanson.html


As an analogy, I find fascinating the MWI vs Collapse debate at LW - Eliezer has concluded MWI wins as an outcome of pure reasoning. No empirical evidence required. It will be (is?) wonderful if (when?) we factorize large numbers quickly and that is evidence of MWI, but as above, it would be (is?) superfluous. If the reasoning offered by Eliezer is coherent (I don't know enough to judge), case closed. Please correct me if I have misunderstood. Thanks

Edit: MWI paragraph added.

Comment author: JoshuaZ 24 November 2012 02:18:45PM 0 points [-]

This doesn't answer the question. Note that no one is arguing that economics doesn't pay rent. The question was whether arguing over '"what is an economy" pays rent or not.

Comment author: Kal 26 November 2012 01:16:16PM 0 points [-]

I misunderstood the Q. In my opinion, yes. There is no other way to get a sound (ie, based on sheer deductive coherence) grasp of the subject. So, attack the issue for yourself:

The first one has a very minor error btw - prob a typo. Good exercise to find it. A friend of mine just pointed it out to me.

Comment author: Multiheaded 25 November 2012 12:00:22PM *  3 points [-]
Comment author: Kal 26 November 2012 01:07:58PM 1 point [-]

Thanks, Multiheaded.

Wonder what a FAI would know about human motivations, dynamics and joys that we don't and thus it chooses differently from the scenario above.

Based on my understanding thus far, this would be a consummation devoutly to be wished. Separating man from man, except where they voluntarily choose to interact. Of course, I likely misunderstand.

Comment author: Alicorn 24 November 2012 05:48:55AM -1 points [-]

Money is one half of every single transaction, in a sophisticated economy.

Economies that have multiple currencies are unsophisticated...?

Comment author: Kal 24 November 2012 06:14:11AM 2 points [-]

I mean that in a monetary economy, one always buys goods and services with money. One side of the transaction is money, whatever it is. Sophisticated here means an economy that has progressed beyond barter and developed a medium of exchange and a store of value, ie, Money.

Important to note separately that multiple currencies are results of legal tender laws. In the absence of such interference, the market would choose one commodity as Money (refer the linked article) and it would not be paper with ink on it. Yes, what could happen is that the chosen commodity is then represented by pieces of paper, but they are simply a more convenient token for the commodity, which is kept in store. And there could be multiple different papers with different names, but they all refer to the stored commodity and are thus economically equivalent and with fixed ratios among them. This economically sound system used to exist - close enough - in the 19th century.

Even now, the various currencies are not different in any meaningful economic sense. Non-US countries use the USD as the reserve for their own currencies. Some also use the Euro, Yen etc. But analyze the Euro and one sees tender laws and gold standing behind it. What stands behind the USD? Laws and legitimacy of the govt enforcing those laws. There is gold, which would be called upon if peple reject the currency as it is. We need to separate the surface phenomena from the economic relationships underlying them.

Comment author: Kal 24 November 2012 05:52:17AM 3 points [-]

I recently read his Fnargl series of posts and the posts on his political journey from Mises to Carlyle and why he is no longer a Libertarian.

I have also read and tried to understand (I probably misunderstand) Eliezer's posts & debates on FOOM, CEV etc.

So, here is a (mostly tongue-in-cheek) scenario combining both Eliezer and Moldbug's ideas.

The FAI (obviously having the powers of the Fnargl-Alien and more) functions as the ideal global Sovereign/Govt - ie, it perfectly enforces all rights of all sentient creatures, functions with an infinite time horizon and protects the planet and all future abodes from other-worldly threats (asteroids and so on).

Whether CEV leads the FAI to the same conclusion, I cannot know. But I sure hope so.

Comment author: pragmatist 23 November 2012 05:37:13PM *  3 points [-]

I will read the Moldbug article, although I have to say I'm not optimistic, given my past experience with his writing. But I do want to disagree on this:

Inferring causality from a time-series of various economic variables is incoherent... Data about effects of economic policy tells us nothing if one has not already pre-supposed causal relationships between certain variables (ie, which variable is affecting which). If one had not done so, how does one know which variables to link with which other one?

I think you are either underestimating the tools at our disposal for causal discovery, or you have an evidential standard for causality that is way too high. Are you familiar with Judea Pearl's work on inferring causation from probabilities? Check out this sweet post by Eliezer if you're not.

ETA: I'm only a few paragraphs into the Moldbug piece and my hackles are already rising. He has already declared that a conscious being must be rational "by definition" and that other peoples' desires are unknowable, also "by definition". I don't think either of these claims are true (assuming the words have their usual meanings), let alone true by definition. This doesn't bode well, but I'll keep reading.

Comment author: Kal 24 November 2012 05:12:35AM 2 points [-]

Thanks. I have read that post by Eliezer before. The issue with monetary economics is the number of variables. Money is one half of every single transaction, in a sophisticated economy. A sound economist's standards for evidence are not any higher than anyone else's, or should not be. It is just that the array of variables is huge and gathering enough data for inferring causality, in the way the post shows, is a pipe dream. Solution: deduction from first principles.

Economists assume certain causal relationships and just screen off everything else. This is so wrong, it is tragic. And its results will be tragic too. The last 5 years were a warm-up act.


Re definitions, the meaning of rational there is that the person acts based on his internal map of the world (this is how Mises used the word way back when and it is part of parlance in Misesian economics). It does not mean what LW thinks it means.

Re unknowable desires, it is a way of saying that economically relevant desires are revealed by economic actions. Demand means being willing and able to pay.

Semantic issue in both cases :)

Comment author: JoshuaZ 23 November 2012 04:18:07PM *  1 point [-]

Inferring causality from a time-series of various economic variables is incoherent. There first needs to be a deductive understanding of what the causal relationships are between economic variables. That is what is meant by "first principles" here - perhaps the disagreement is semantic.

If I replace "economic variables" with "astronomy" what part of this sentence changes in implication? Why is this incoherent for some fields? The level of rigor cannot be the only difference: Physics and astronomy have become more rigorous over time, not by discussing first principles, even when moving from Aristotle to Medieval physics to Newtonian physics, but rather by adopting principles based on the empirical data. The same goes for the switch from Ptolemaic systems to Copernicus and then Kepler. It is the empirical data that matters, the apparent time-series of planetary and stellar motion.

Thus, the first principles analysis: What is an economy?

Does arguing over this definition pay any rent?

Comment author: Kal 24 November 2012 04:55:00AM 1 point [-]

Re the Q about rent-paying: Yes. You should rightly be skeptical now, but please read the Economics sub-section at Moldbuggery with an open mind. Or if you prefer, start with Rothbard's 'Man Economy and State'.

The reason I linked to Moldbug's articles in this thread (indeed, this thread being only reason I signed up instead of lurking) is cos I have seen it said in LW that Misesian economics is incoherent, anti-empirical etc. With respect, people who say that are making a mistake. But what is germane is not the mistake, but the financial effect it can have.

Economics - of all subjects - pays rent. Maybe not in the short term, but eventually and especially when the central banks of the major economies are doing effectively insane things.

Comment author: JoshuaZ 21 November 2012 09:23:25PM 5 points [-]

The leap from "controlled experiments are not possible" to "one has to deduce from first principles" is huge and unsupported.

Yes, whenever people try to make this sort of argument I have to wonder how they think we should do astronomy.

Comment author: Kal 23 November 2012 03:39:26PM 1 point [-]

Please see my reply above to pragmatist.

To add a bit, the rigor in monetary economics today is so far behind physics, it is not fair to compare the two subjects. It is an insult to Physics.

Comment author: pragmatist 21 November 2012 09:20:47PM *  12 points [-]

The leap from "controlled experiments are not possible" to "one has to deduce from first principles" is huge and unsupported. The results of controlled experiments do not exhaust the available empirical evidence by a long shot. We have a lot of data about the effects of monetary policy from around the world. True, inferring causality from this data is not nearly as straightforward as inferring causality from a randomized controlled trial, but it's still a lot more reliable than deduction from first principles, I would think.

Think about how your argument sounds when applied to cosmological theories about the very early universe. We have a number of different theories that cannot be simultaneously right, and we cannot conduct controlled experiments. Would you endorse deduction from first principles in this instance as well?

Comment author: Kal 23 November 2012 03:36:26PM 0 points [-]

Inferring causality from a time-series of various economic variables is incoherent. There first needs to be a deductive understanding of what the causal relationships are between economic variables. That is what is meant by "first principles" here - perhaps the disagreement is semantic.

Data about effects of economic policy tells us nothing if one has not already pre-supposed causal relationships between certain variables (ie, which variable is affecting which). If one had not done so, how does one know which variables to link with which other one? The causality which is being claimed to be derived is actually assumed or it is a result of the data-mining effect.

More importantly though, I suggest (for what it is worth) that you ignore my comments which are replies to others' comments and try the original article. It repays inspection.


I have no expertise in cosmology. So I have no opinion on that which is worth stating. The only comment I have is that Physics has causal relationships clearly defined and known. The different theories thus could make different testable predictions about what we should expect to see in the here and now.

In economics, the different schools (more accurate to call them ideologies) do not agree on basic causal relationships between variables. In more than one case, one of the main tenets (of a popular ideology) is plain, flat-out incorrect, and sometimes definitionally misunderstood. I know it may be hard to believe that monetary economics is that shaky a subject, but it is. The last hundred years have been a regression when it comes to monetary economics. Thus, the first principles analysis: What is an economy?

View more: Next