Friday, May 22, 2009

Human Action: Chapters 1-2

Summaries for chapter 1 and chapter 2 are available from the Mises Institute. If you were constantly looking up words (or trying to figure out what that latin phrase meant), check out Mises Made Easier.

Rather than summarizing these chapters yet again, I thought it would be more fruitful to discuss some of the questions I had while reading. Feel free to add your own questions and comments; hopefully we can get a good discussion going.

The biggest question I had while reading these two chapters concerned subjective probability. On page 23, Mises makes it clear that individuals are capable of acting in accordance with (objective) probabilities.
Sometimes we succeed in acquiring a partial knowledge so that we are able to say: in 70 per cent of all cases A results in B, in the remaining cases in C, or even in D, E, F, and so on. In order to substitute for this fragmentary information more precise information it would be necessary to break up A into its elements. As long as this is not achieved, we must acquiesce in what is called a statistical law. But this does not affect the praxeological meaning of causality. Total or partial ignorance in some areas does not demolish the category of causality.
Then, on page 37, he explains that an individual's beliefs about reality matter.
A peasant eager to get a rich crop may—according to the content of his ideas—choose various methods. He may perform some magical rites, he may embark upon a pilgrimage, he may offer a candle to the image of his patron saint, or he may employ more and better fertilizer. But whatever he does, it is always action, i.e., the employment of means for the attainment of ends. Magic is in a broader sense a variety of technology. Exorcism is a deliberate purposeful action based on a world view which most of our contemporaries condemn as superstitious and therefore as inappropriate. But the concept of action does not imply that the action is guided by a correct theory and a technology promising success and that it attains the end aimed at. It only implies that the performer of the action believes that the means applied will produce the desired effect.
If individuals are capable of using probabilities and their perception of reality (which is not necessarily equivalent to reality) is the driving force behind purposeful action, why doesn't it follow that their perception of the probability of a given event is more important than the actual probability that the event occurs? Or, put differently, if individuals can be systematically biased about "theory" and "technology," why can't they be systematically biased with respect to the outcomes they anticipate?

Bryan Caplan makes a strong case for subjective probability. Walter Block counters; as does Jorg Guido Hulsmann. Caplan responds.

Next week, we will cover chapters 3-7 (again, roughly 70 pages). Happy reading!

17 comments:

Justin Ross said...

If his book was written this year, the same word for word, it would very much read like a critique of behavioral economics. By Mises definitions, every action is rational by definition, which is probably not going to be very appealing to many.

Lets keep him in the modern box, for a moment, and instead try and sort out the perceptions vs objective realization problem. We know from behavioral economics that people will treat "1%" differently than "1 in 100." Objectively these are the same, but peoples perceptions of them are indeed different.

My reading of Mises leaves me with the impression that this is ok for his views on reality. That people my preceive the two figures differently but still use them in the same rational construct for their behavior. Rather than compare "X%" to "X in 100", they still are able to rationally use "X% vs Y%" and "X in 100" vs "Y in 100".

Justin M Ross said...

Mises’ view of rationality and happiness summarizes the foundation of libertarian philosophy, particularly on p. 14:


What makes a man feel uneasy and less uneasy is established by him from the standard of his own will and judgment, from his personal and subjective valuation. Nobody is in a position to decree what should make a fellow man happier.However, his critique of the opponents of this view is rather shallow. Notice that we should all be able to agree that the statement taken literally is also false. I am in position to decree what is best for my toddler. He reiterates this same point without acknowledging the deeper criticism, as the real argument between libertarians and their critics usually occurs on the parameters of the statement, i.e. what constitutes a “man” that can be considered capable of entering into a contract? Libertarians will acknowledge that children and mentally disabled cannot make a contract, but how do you define such a group? Libertarians actually have a higher tolerance for the range parameters and a limited number of dimensions where they’d be willing to strip them of self-determination. On the other hand progressives, for example, have a much narrower view of who is capable of making their own decisions. Sticking with progressives, the outcomes of one’s personal choices can be used as evidence against the case that they can make their own decisions. “Their income is low? That’s because they are witless and exploited...we need to bargain on their behalf.”

Mises, as thus far, has not really addressed how those parameters can or should be defined. What age makes an adult? What IQ is mentally competent? What dimensions are up for consideration? Perhaps it is unfair to ask him to do so.

Will Luther said...

I am glad you also read these chapters as "a critique of behavioral economics." I agree with your conclusion that "people may perceive the two figures differently but still use them in the same rational construct for their behavior." However, modern Austrians entirely reject the idea of subjective probability/Bayesian updating. I have to read Block's and Hulsmann's critiques once more. But, so far, I am not persuaded by the Austrians on this point.

Anonymous said...

I would use some caution in equating "modern Austrians" with Block and Hulsmann. They are the fringe within even the Mises Institute when compared with people like Roger Garrison. And, the Mises Institute tends to be more fringe than 'Austrians' who attend the Society for the Development of Austrian Economics meetings. I'm sure many of these people, and even some people at GMU, don't "entirely reject" any sort of Bayesian updating.

Will Luther said...

Anonymous: I understand where you are coming from with respect to the SDAE, but Block and Hulsmann are regularly published in the QJAE. To some degree, this makes them authorities on Austrian Economics.

With respect to the Mises Institute, my impression is that Garrison is on the fringe (I could be entirely wrong about this, of course. I have never visited the Mises Institute. This is just the impression I got from reading the articles critiquing Garrison's work in the QJAE.)

With popularity contests out of the way, I still think most Austrians would reject the notion of Bayesian updating (maybe I should reconsider the word "entirely"...). This was certainly the case when I brought it up over dinner with Pete Boettke, Pete Leeson, and Richard Ebeling about a month or so ago.

Anonymous said...

Luther, what do define as subjective probability and Bayesian? What did your advisors oppose about it?

jb said...

You are in a position to decide what is best for your toddler, but I doubt that you are in a position to truly know what makes him happy at every juncture. If you did know such a thing, you could establish your own mint by providing such information to bewildered parents! The conundrum, as you aptly state, is exactly when one become a truly conscious being, as Mises points out the reason we cannot remember our early years is because we have nothing to recall because our minds have not sufficiently developed. It seems like a scary Orwellian decision to make for a libertarian, if applied legally.

Mises seems to attempt to be very precise in his logic, but reached a barrier which is outside of the scope of his argument, and possibly left to groups willing to make a subjective determination from his general guidelines. I find no problem with his assertions at this point, and think that parental decisions do not contradict his words to this point.

I could be wrong.

jb

jb said...

I simply think that Austrians reject the quantification of human decision-making. Thomas C Taylor wrote in "An Introduction to Austrian Economics", "Uncertainty relates to situations that are unique; each situation is a case in itself as opposed to being a member of a class or large number of homogeneous events or circumstances. Uncertainty is not numerically calculable because of the lack of sufficient past experiences relating to the particular set of circumstances being considered."

I could be wrong.

JB

Justin Ross said...

I think you are right, JB, that my hope for him to address those parameters is probably unfair.

As a Bayesian, I am surprised to learn there is debate among Austrians over the subject. There is a heavy debate between Bayesians and frequentists that I am in no way qualified to contribute to, but as you can probably guess I find Bayesians convincing enough to use their work.

If anyone is following along and wants a simple clear intro to Bayes: http://yudkowsky.net/rational/bayes

Will Luther said...

Anonymous:
JB gets the main point. Bayesian updating is too mechanistic for Austrians. It leaves no room for radical uncertainty. In other words, it is not that you hold some events with very, very low probabilities (as Baysians would suggest) but that you don't know that you don't know what you don't know.

I encourage you to read Caplan, Block, and Hullsmann (which I linked to in the original post) to get to the details of the debate.

Suzie said...

I’m slacking guys (Will didn’t your first post state it was only to be 40 pgs. a wk or was I just dreaming that as I fell asleep next to my computer?)

For the First Two Chapters here are some questions/thoughts I had (I put a disclaimer on this that I am not an economist connoisseur yet so this may be really basic stuff, but that will be good for any readers who may be at my level)

From pg 22: “In a world without causality and regularity of phenomena there would be no field for human reasoning and human action.” I find this quote especially interesting because it matches well with psychology/social work theory. It is noticed that children (and most people in general) thrive on consistency. This enables them to understand what comes next in a normal daily routine and therefore they can learn how to plan accordingly. When chaos enters a family, children may not know whether their parents will hit them or applaud them for the same behavior. It may make them anxious about any action because they do not know what to expect, and therefore they do not know what action will bring the ends that they want to produce (which could be not to be beaten or even to make mom happy). However abused kids do try to find ways to explain their parents actions and find ways to cope/stay alive, but it is possible if life for everyone was completely irrational our ability to reason may not have developed.

Suzie said...

On page 55: “There are in the field of economics, no constant relations and consequently no measurement is possible.” Is this implying that we are never going to be able to measure economic realities and have it have implications for the future because the current state is ever changing? Is the only thing we obtain from studying economics is what was, not what is and because of this, we cannot learn from what was because it is now different? If this is true, why then do some economists give recommendations for policies (if their knowledge is only relevant in the past)? And then how is there a law of economics (as the quote below states) if everything is changing? Why are we reading this book if today is completely different than when it was written?

I thought it was interested no one commented on this beauty of a quote pg 67: “In fact, economic history is a long record of government policies that failed because they were designed with a bold disregard for the law of economics.”

Suzie said...

Pg 57: “… even with the same perspective many disagree with relevance of facts. As far as understanding aims at relevance to each factor, it is open to the influence of subjective judgments. Of course these are not judgments of value, they do not express preferences of the historian. They are judgments of relevance.”
To me, judgments of relevance sure sounds a lot like subjective values. So if a statistical relationship shows that the percentage of people in a state who are High School (HS) dropouts is negatively correlated to the percentage of people who have health insurance in a state, JRoss and Will may have different explanations for why HS dropout percentage is a relevant factor in the relationship. Wouldn’t this be their subjective perspectives on the relevance of the factor? This is then how different than their ‘value’ of HS education? Wouldn’t this value go into their determination of the relevance?

On page 60:”But it is logically impossible to make the membership of a class or type dependent on an average.” My first thought on this was what about logistic modeling to predict if someone would be within a group/category etc. Do you think that this has changed due to the progression of econometrics or are these ideal categories still impossible to define based on math (granted, the math does not always predict correctly either)? Also, I guess his point would be that to even create the model, you need data with this ideal confirmed for the observations.

My final thought on a first read was this:
Based on pg 7 “Science does not give us absolute and final certainty, it only gives us assurance within the limits of our mental abilities and the prevailing state of scientific thought. A scientific system is but one station in an endlessly progressing search for knowledge… It merely means economics is a living thing and to live implies both imperfection and change.”
With this being said how do you feel knowing that your current tools and base of knowledge may be totally reconstructed one day due to a new expansion of thought or understanding? Is it all in vain? Or do you believe that we will do the best we can with what we have and hope to develop some understanding through economics even if the current tools are not 100% accurate. What happens though if something comes to disprove it all, just as our current scientific understanding made many superstitions irrelevant?

Justin M Ross said...

I'm glad Suzie got us off of subjective probability...

Mises critique of econometrics and statistics here plays up one of the more prevalent stereotypes about Austrian economics...that they don't/can't do Math and Statistics or otherwise don't believe in it.

I also had trouble putting together what Mises wants me to take away as his point of view on this subject. My impression is that his critique is that:

1) Correlations can only help confirm our prior theory in a general directional sense, but not in the sense of a fixed quantity. I guess I am tempted to say "he will accept empirical work that the effect is positive and statistically significant, but he'll never believe that the beta = 2." So we can establish as a law that as price goes up, quantity demanded goes down, but we can never say with any certainty by how much.If this is the case, then Bayesian econometrics should be more appealing, not less, to Austrians.

Nevertheless, if an alien were to watch my behavior on a regular basis (I'm thinking of the TV show Blossom, here) I'm sure the would infer causality very differently from reality. Surely they would regard the small ones in our household to be the masters and the big ones to be their servants. They would also likely think the bus that stops outside my house at 7:30 am causes me to eat Cheerios.

2) The limitations of empirical work is that it will likely always be case/time specific, so it is better to rely heaviest on theory.
...........
It seems to me that the Austrians focus much more heavily on the complexity of the market process. I would even go so far as to regard it as an infinitely complex process. In which case it seems to me that the Mieses critique of econometrics should be something along the lines of not being able to draw a sufficiently large random sample from something that is infinite in size.

Anonymous said...

Will, it seems like traditional statistics would be even more "mechanistic" and leave just as little "room for radical uncertainty." Is this true? I was under the impression that Bayesian was more dynamic. Sure it doesn't include radical Knightian Uncertainty, but that doesn't make it useless, does it?

As for the Austrian view of econometrics, I think it is just as useful as any other form of empiricial research, what Mises terms "history." The point, however, is that doing "history" (in any form) cannot determine theory. The world is to complex and dynamic to simply look at it. Hence, a priori theorizing is needed to to establish certain propostitions.

Also, many of the critiques of metrics are appropriate for the methods of the day, but have become less relevant because of advances in the field. Bob Subrick has an interesting paper on this titled something along the lines of "What old Austrians need to stop saying" or something like that.

Will Luther said...

Suzie: I am glad you are joining us.

"I find this quote especially interesting because it matches well with psychology/social work theory. It is noticed that children (and most people in general) thrive on consistency."

I think Mises is making a much smaller point here. He is merely saying that one cannot act rationally if particular actions are not correlated with particular outcomes. Imagine that turning on the faucet is completely uncorrelated with water coming out of the faucet. You can recognize your thirst and grab a glass from the cabinet, but there is no action you can take to fill the glass with water. Fortunately, we live in a world of causality and regularity of phenomenon; rational action is possible.

"Is this implying that we are never going to be able to measure economic realities and have it have implications for the future because the current state is ever changing?"

Justin nails this, imho. Positive and statistically significant, but the beta value is irrelevant.

"With this being said how do you feel knowing that your current tools and base of knowledge may be totally reconstructed one day due to a new expansion of thought or understanding? Is it all in vain?"

I believe our understanding of the world grows as ideas come in conflict with each other. I frequently refer to this process as the great battle of ideas. And I believe it is the only way to get to truth. When ideas battle it out, the rough edges are worn down. The working theories are refined. We retreat, when necessary, and rework our positions. After awhile, we may be arguing from a completely different position. But it was a gradual process that got us there. And our ideas are much stronger because of it.

Anonymous: Traditional statistics is much more mechanistic. However, it is only useful in a relatively small number of situations where all objective probabilities are known. So, at the blackjack table, where the odds are fixed, individuals will use probabilities mechanistically. But in the supermarket, and our dynamic world more generally, we do not have fixed objective probabilities and all events are unique.

Anonymous said...

buy facebook likes
get facebook likes

http://www.libgig.com/node/2171 http://www.huwy.eu/ie/result/internet-0
buy facebook likes buy facebook likes facebook likes
I already unistalled my avast anti virus software I even use and unistaller software to delete it and I also tried using control panel...but it is still ther...everytime I started the computer a window will open and said avastblahblah has stop working...and that makes my computer to work slow.. any Idea how to stop this

1000 facebook likes buy facebook likes [url=http://1000fbfans.info]get facebook likes [/url] 1000 facebook likes