?

Log in

No account? Create an account
Pattern Recognition
June 30th, 2012
08:33 am

[Link]

Previous Entry Share Next Entry
Sketch of a Right-Libertarian Argument for Economic Redistribution
Disclaimer: The following contains some assertions about how classical liberals/libertarians and Marxists think about economic justice. These are only intended as very crude simplifications of the intuitions usually brought to bear in Internet message board rants, angry screaming at NPR during long car trips, and OWS encampments. Your own beliefs are, of course, infinitely more subtle.

Other Disclaimer: This contains some handwavy math. Sorry about that.

Suppose an economic actor (a real person, or a legal person, i.e. a corporation) takes an economic action (by itself, or interacting with others) and has an economic gain or loss. What happened there? There are two traditional explanations, and a third that maybe doesn't get enough attention:

Classical liberals, e.g. American conservatives/libertarians, tend to assume that economic gains and losses are earned. Market outcomes, in the absence of force or fraud, are axiomatically fully just.

Socialists/Marxists tend to assume that economic gains under capitalism not attributable to labor inputs are stolen, fully attributable to exploitation.

Finally, there's plain dumb luck.

(Modern finance uses a somewhat more sophisticated model when they speak of the "alpha", the portion of a gain that is attributable to the actor's real merit--whatever the hell that is--so their model is alpha ~ merit and (1 - alpha) ~ luck.)

With a little thought, I hope the reader will concede that most economic activity involves a mix of these factors: actors take more or less constructive action, influenced by luck, and sometimes gain or lose from unjust, exploitative, criminal activity (by them, or by others acting against them).

We could then describe the actor's wealth as a sum of three distributions: merit + luck + exploitation. Getting a concrete breakdown of how much wealth is attributable to which factor is pretty much impossible; which outcomes you want to attribute to merit and which to exploitation, in particular, depend on your theory of justice. OK. For this discussion, though, that doesn't matter. Just consider merit, luck and exploitation as some unknown random variables, and consider the distribution of the equation:

Wealth = Merit + Luck + Exploitation.1

In particular, look at the tails of the wealth distribution!

Someone at the high end of the distribution is, in all probability, a triple winner: high merit + high luck + high exploitation. Likewise, someone at the low end is most likely a victim of unwise actions, plus terrible luck, plus victimization. Ayn Rand is wrong, and so is Occupy Wall Street.

In a free market we should value clear economic signaling -- people should profit from their actions in a clear and predictable fashion. Nevertheless we can conclude, statistically, with a high degree of confidence, that people at the high end don't merit everything they accumulate and those at the bottom don't deserve all the shit piled on them.

And...so that's basically it. Trying to unpack the 'true' value of alpha is pretty much impossible. We have a huge, expensive justice system that almost nobody is happy with, so we know vividly just how bloody hard and expensive computing the exploitation factor is, even if you accept the existing standard of justice.

But we can say with some confidence that even within a strict free-market framework that there is some factor 1 - alpha that, if confiscated from the wealthy and redistributed to the poor, increases overall justice without truly punishing merit, if only because dumb luck is always a factor.

How you accomplish that redistribution in a just way is a whole other set of problems, of course...



1: Interestingly, both the classical liberal model and Marxist models reduce to something like:

Wealth = Merit + Exploitation

The difference (to be very crudely reductive) is that classical liberals think Exploitation = Force + Fraud and everything else is Merit, while Marxists think Merit = Labor and everything else is Exploitation.

Tags: ,

(59 comments | Leave a comment)

Comments
 
[User Picture]
From:sealwhiskers
Date:June 30th, 2012 03:39 pm (UTC)
(Link)
Hm, this seems like a fairly brave post, even if carefully phrased. I don't disagree. But then of course I am always an advocate for mixed systems, preferably leaning in certain directions, but still.
[User Picture]
From:st_rev
Date:June 30th, 2012 03:44 pm (UTC)
(Link)
Brave because the libertarians will come after me, or brave because the communists will?
[User Picture]
From:fengi
Date:June 30th, 2012 03:57 pm (UTC)
(Link)
I like it. It's much better than my objectivist argument for communism which is that if rational individual self interest is the ultimate good, it's in one's rational self-interest to band together with like minded people and share.
[User Picture]
From:st_rev
Date:June 30th, 2012 03:59 pm (UTC)
(Link)
There's an offhand remark in one of Ken MacLeod's early novels to the effect that if Marxism is to succeed, it has to be able to compete in the free market. Something that stuck with me.
[User Picture]
From:caladri
Date:June 30th, 2012 04:38 pm (UTC)
(Link)
I suppose a useful extension to this is to ask whether it matters if some market activity is coercive. In a truly free market nobody would be able to coerce anyone into paying more than they need to for something (or paying for something they don't need), perhaps, but that seems unattainable, or at least the methods for attaining it are invisible for me. We can take corrective action for that, in the form of coercive models of redistribution that probabilistically will mostly take the cream of exploitation and redistribute it. That in and of itself would be a market action, if a collective market action. People banding together and forming a government that participates in the market isn't that different from people banding together and forming a corporation that participates in the market. Libertarians get mad when those redistributive actions are not voluntary, which is always reduced to images of the government holding a gun to a hard-working meritocrat. But are usually suspiciously-silent on the exploitative and coercive forces that non-government actors may profit from within the market sphere. Unless those forces tend to be corrective for exploitation, i.e. are poor people robbing rich people. So is it really the methodology that's the problem, or the purpose of the exploitation? Why do monopolies need fewer regulations (i.e. restrictions on wielding market power) that governments? Why do upwardly-directed economic violence and state-driven market activity need to be eliminated, but not the same done by non-government corporations?
[User Picture]
From:fengi
Date:June 30th, 2012 05:11 pm (UTC)
(Link)
Good points. I wonder if the issue is that libertarians - or the version we're discussing here - aren't egalitarians even if they seem to be very similar on the surface.

On a side note, one thing this discussion makes me realize is every political philosophy and theory is prone to distortions made to accommodate one's own preferences and preferences, libertarianism is one where the convenient exceptions stand out rather easily.
[User Picture]
From:nihilistic_kid
Date:June 30th, 2012 05:12 pm (UTC)
(Link)
Aw, the right-libertarians would just switch from utilitarian claims to natural law claims.
[User Picture]
From:st_rev
Date:July 1st, 2012 02:12 am (UTC)
(Link)
I am a utilitarian right-libertarian. Natural-law libertarians are religious fanatics. Not much you can do with them.
From:(Anonymous)
Date:July 1st, 2012 12:27 am (UTC)
(Link)
I am pretty sure that the people on the high end of the distribution would explain to you that -- in their case -- the weighting of the luck component is extremely small, and the people at the other end would explain somewhat less patiently that the weighting of the merit factor is extremely low on their end. It's extremely hot in my office right now, but I'm pretty sure your point includes that these claims, even if true, are irrelevant if you get far enough out in the tails? I'm not sure about that, see previous remarks about heat.

There must be some study out there explaining that people attribute negative outcomes to "luck" and positive outcomes to "merit".
[User Picture]
From:st_rev
Date:July 1st, 2012 02:22 am (UTC)
(Link)
I'm pretty sure your point includes that these claims, even if true, are irrelevant if you get far enough out in the tails?

Yeah. Here's an extremely simple model of what I'm talking about: Say I have three six-sided dice, labeled 'merit', 'luck' and 'exploitation'. Rolling three dice and adding the results gives a bell-shaped distribution of outcomes between 3 and 18.

A number in the middle of that distribution can be rolled in a lot of different ways. If I get a 10, that could be 6 merit, 1 luck and 3 exploitation, or it could be 1 merit, 3 luck and 6 exploitation, for example.

But if I roll an 18, I MUST have rolled 6 merit, 6 luck AND 6 exploitation. A 3 must mean 1 merit, 1 luck and 1 exploitation.

Disclaimer: Economic processes don't behave like dice.

There must be some study out there explaining that people attribute negative outcomes to "luck" and positive outcomes to "merit".

Yeah, it's a theme in Nassim Taleb's work. Particularly when you're in a position to take benefits from positive outcomes and foist off negative outcomes on others.

Edited at 2012-07-01 02:55 am (UTC)
[User Picture]
From:cubes
Date:July 1st, 2012 12:31 am (UTC)
(Link)
Your assertions regarding the tails are perfect.

Your last few paragraphs are also, alas, spot on.

I think a lot more arguments than people realize are around who and how much and which way it's done rather than the basic fact that it should be done.
[User Picture]
From:st_rev
Date:July 1st, 2012 03:11 am (UTC)
(Link)
My concerns are very much with the how, and I'm fairly pessimistic. In a lot of ways, it seems to me that the actually existing welfare-state apparatus may be worse than no safety net at all. A straight guaranteed basic income would be a lot better, I think.
[User Picture]
From:richardthinks
Date:July 2nd, 2012 09:01 am (UTC)
(Link)
just writing to applaud. Simple, clear, necessary, somehow missing from all literature up to now.
[User Picture]
From:st_rev
Date:July 2nd, 2012 09:06 am (UTC)
(Link)
Thank you, that means a lot. I have a lot of thoughts about this constantly running in the back of my mind but it's really hard for me to put them down on the page. I was afraid this post would sink like a stone
[User Picture]
From:twoeleven
Date:July 2nd, 2012 03:51 pm (UTC)
(Link)
In order to comment in a sensible way, I'd like to be clear what we're talking about. So...

What's a "right-libertarian"? What do they believe? Specifically, what do they believe about "justice"? (If it would help, replace "they" with "you" in the previous sentences.)


Statements of bias:

I'm always suspicious of "* libertarians" and "* anarchists", because my experience with their beliefs has frequently led be to concluded that either a) they're walking examples of magical thinking and/or b) their beliefs can be summarized as "I want a very limited government, except when I want my government to do otherwise-indefensible things to people I don't like.". I don't expect either of these from you, however, I though you should know I've got a wee bit of bias about the subject.

The dodgy math is already making me twitch. Sorry. For a start, "luck" and "merit" are not orthogonal, which causes significant problems all by itself. The catchphrase that successful people "make their own luck" is partially true, and in fact, I heard a lecture on that subject last Friday.


Coincidentally, I was gonna post on a similar topic from a different point of view.
[User Picture]
From:st_rev
Date:July 2nd, 2012 04:27 pm (UTC)
(Link)
What's a "right-libertarian"?

It's a meaningless word in the title of the post, not actually relevant to the argument per se. The argument does assume some classical liberal norms as framing, e.g. it assumes that markets and economic actions actually exist and aren't some kind of pre-postmodern word game or a demonic evil that must be purged.

What do they believe? Specifically, what do they believe about "justice"?

I tried to be explicitly agnostic about theories of justice. As nihilistic_kid noticed, the argument is utilitarian, and would not be acceptable to a natural law theorist or someone who otherwise differs on some fundamental axioms.

I'm not sure what to say about your suspicions. If it offends your prejudices less, feel free to think of me as a Hayekian.

For a start, "luck" and "merit" are not orthogonal, which causes significant problems all by itself.

Division of outcomes into merit vs. luck vs. exploitation depends very much on your concept of justice in theory, and is absurdly difficult to determine in practice, although a great deal of work has been done in the area of finance on e.g. calculating the alpha value of fund managers. If you like, you can subsume all risky but deliberate economic behavior1 under 'merit' and only count purely happenstance events, e.g. residual losses from insurable risks, as 'luck'. (You'd be wrong, but that's a much more complicated argument.) The argument here still works, although the 1 - alpha factor may be vanishingly small.



1: Risky but deliberate actions where you accept all outcomes, that is. A fund manager who is positioned to profit from fund gains but insulated from losses may be benefiting from, in effect, a fradulent derivative instrument; the whole issue of the posterior fallacy also comes into play here.
[User Picture]
From:luagha
Date:July 8th, 2012 10:52 pm (UTC)
(Link)
I had something more pithy when this came out but I was unable to make a comment - some sort of lj bug of the moment, new user interface thing. Now I've lost it.

What is coming to me now is that using 'sums' is too simple. I use as my vague counterexample the Warren Buffett of the 70s-80s-90s versus George Soros.

Warren Buffett does research and follows a long term buy and hold strategy that is simple and highly effective. It strongly minimizes luck. It requires a sizeable investment base to make yet more money. It takes advantage of long term capital gains taxes. He is top end wealth, high merit, low luck, moderate to low exploitation. (He becomes moderate-to-high exploitation in 2008 with his connection to the Obama regime and a big inflow of crony capitalism.)

George Soros is an expert trader who trades on inside information. He uses spies in government organizations to inform him of their slow moving policies - like a peg to a certain value. Then he devises trades against those policies because he knows what move his opponent in the trade has committed to take. (This is how he famously 'broke the bank of england.') Also requires sizeable investment base. Top end wealth, but moderate merit, low luck, high exploitation.

So simple sums are not enough. There are other factors or multiplicative/exponential/factorial factors.

It is a very good general argument for progressive taxation of some kind.

[User Picture]
From:st_rev
Date:July 9th, 2012 01:40 am (UTC)
(Link)
Yes, the factors aren't actually additive, but as I commented above to twoeleven, the argument goes through if for 'luck' you only count pure happenstance.

The argument also weakens somewhat if you assume that the factors are individually fat-tailed; it's possible to construct an interpretation under which the richest can only be assumed to be exceptionally meritorious OR lucky OR exploitative.
[User Picture]
From:fuhchee
Date:February 15th, 2016 03:19 pm (UTC)

"all probability"

(Link)
"Someone at the high end of the distribution is, in all probability, a triple winner: high merit + high luck + high exploitation."

What is your evidence for "all probability"? You're being a bit sneaky, using the "+" sign as if it were additive, whereas you mean "and". You would need to support each term of that AND, not just the sum.
[User Picture]
From:st_rev
Date:April 21st, 2016 11:07 pm (UTC)

Re: "all probability"

(Link)
I am, literally, using the + sign to mean additive: W = M + L + E.

Formally, I'm asserting that, given that X has wealth W, and given no other information, one can make inferences about the distributions of M, L, E, and if W is high, that tells you M|W, L|W, E|W are shifted upward. This is just arithmetic.

Edited at 2016-04-21 11:13 pm (UTC)
Powered by LiveJournal.com