A new book: [A Treatise on Sciences]

I recently published a new book, called [A Treatise on Sciences]. It is for now self-published in Amazon, as I felt a quick publication is more of priority, and I can later re-publish. Plus, I wanted full control on the contents of my book, especially its four-chapter structure on difference sciences. 

The current plan is to publish chunks of the book as a paper in some journal, but I do not know how the future will unfold. 

The link is on https://www.amazon.com/Treatise-Sciences-Bryce-M-Kim/dp/1980370273/, ISBN-10: 1980370273, ISBN-13: 978-1980370277 . 

The book is about foundations of different sciences - though while the book discusses methodology of different sciences, it is not really about methodology. Thus, it is about foundations in sense that it discusses important conceptual questions, than methodological questions. 

Physics, economics, statistics and philosophy/political science are four covered sciences of the book. Out of these four, in terms of scientific contents, the first two are of importance. In terms of overarching vision, the latter two chapters (statistics and philosophy) are more important.

I may go over the contents of the book, but for now I am going to be fine with writing the fact that I have a book.

Film [Real] final "Ballet" Scene

 Obviously, [Real] (2017) is one of the worst Korean films - note that I used "is" instead of "is considered as". I cannot find hard to believe that any person would think this film is on par with some other films.

 It is true that [Real] has few interesting bits. If anyone comes to realize intentions of director-writer, then there are things that make sense. Especially how the writer was managed to "equalize" two characteristics physically (wounds, etc.) and psychologically. Except, these interesting details are not going to do any improvement on executions of this film.

 There are some people who say that the film needs more explanations - but even when cutting out some unnecessary scenes, especially involving women (this film should really be condemned for its portrayal and "character use" of women), these more explanations would make the already-very-boring film a yawn inducer.

 That is, this film is simply one of worst because it causes boredom in an unexpected way - its only accomplishment.

 And of course comes this ballet (?) fight scene which causes small laughter for some people - not because the scene is funny but because the scene appears out of nowhere. But ironically this scene actually is the best scene of the film, if we decide to ignore the story plot. True, this scene... is horrible, but then if the film consisted only of this part, at least it works as a nice dance...... I don't know how to provide any meaningful thing to this scene, but then at least it captures your mind for its "ordinary" grotesqueness. (The film is distastefully grotesque in general, and this scene is the only scene that is not distasteful.)

 Anyway really a non-serious just-for-fun post. And the film overall had the taste of being written by someone in extreme stress and wanted to go crazy. I know this because I sometimes felt like that.

----

Addition: Action scenes are beyond terrible. You will see what that means when you watch the film. Plus, for this "but the messages of the film were great":

The messages are (I used "is" instead of "are" intentionally below):

"Who is you - the spirit that occupies your body, or the spirit that used to own your body? Who is you - the one that you wanted to be, or the one that you actually are?"

OK. A proper response to this is: SO WHAT? In philosophy, you can think about free will, Nietzsche and so on, but this film really has nothing to do with these things. The film simply hovers around the messages and does nothing. It is film's grotesqueness that makes you think that the film has unique messages, when it really has nothing of that sort. It simply has no message and no story. You just have a crazy person dancing at the end.

Statistics: Scientists and Engineers

Again, a short memo.

You can pretty much get what I am going to say here from the title: I really think we should think of statistics in two frames: science perspective and engineering perspective. And this, in a way, is inspired from Greg Mankiw's nice writing about how we should view macroeconomics in these two frames (see http://www.nber.org/papers/w12349 ).

I had a 180-degree turn on how statistics should be (mentioning in case anyone knows of my past thoughts) - and I now believe that statistics should more become like machine learning (more on how machine learning is different from statistics later) - or at least as a matter of "science."

That is, I do think that how machine learning thinks of statistical inference - basically learning algorithms - is the right way to do statistics science.

But not all sciences are complete, and some tools being developed scientifically often are incomplete and not ready for use. This is why we need engineering tools. In statistics, this tends to be human modelling intuitions and tips built from experience confronting reality. When one science is developing, far from completion, there are engineering tools that seem to work in reality and yet feel suspicious in terms of existing scientific frameworks. For example, Dirac's hole theory was replaced with a different scientific understanding, and yet in engineering the idea of "holes" is sometimes still taken up for convenience. With hindsight, we can consider Dirac's hole theory as an engineering tool to describe reality even if it is incomplete.
Similarly, we can even understand "Standard Model" in physics as an engineering model of the future, once we get to understand how a unified physics theory works.

To go back to statistics, there is the question of "how statistics should be" (science) and "most useful pictures statistics can provide" (engineering). There are times when these tend to go differently. Thus, it may be true that statistics should more be about learning algorithms, but for current understanding of learning algorithms, more traditional statistical approaches may be the best option available (though we can understand them also as particular learning algorithms).

------------------------------------------

Addition: it is often said that machine learning tools often tend to look at data and have disregards to structural models. But this is not necessarily the case. Think of the case on how machine learning tools would incorporate knowledge of classical physics, which can be said to operate in our fairly non-relativistic world (that is, we can live as if the world is non-relativistic most of the time). Simply make a meta-learning algorithm being used to choose learning algorithms that make use of classical physics knowledge that analyze data so that analysis never does not go in conflict with classical physics. It is actually traditional statistical approaches that have harder times incorporating these established knowledge. And in terms of checking established truths, machine learning approach is much better in dealing with data inconsistent with established truths, but this is a topic for another day.

(One can also think of typical machine learning problems. Many of these problems can be clearly stated - like "winning strategies in Go" - and thus one can think of learning algorithms that work best for these particular classes of problems, instead of general ones. One can go pure theoretical and derive an optimal learning algorithm using characteristics of these problems. Or one can go somewhat empirical and think of what has so far worked out best.)

A short memo on the state of economics

A short memo that may possibly lead to a new essay on the state of economics.

While a lot have been said about the state of economics, there has been a hole that very few, if not none, notice.

That is:

 Arrow-Debreu-Mackenzie general equilibrium theory (ADM theory), restricted to conventional understanding, cannot actually be the foundation of modern economics done so far, if not for future. This is true even when we leave out issues like behavioral economics. (It has been claimed that much of economics has, as its foundation, ADM.)

One may argue that game theory forms a more of general foundational theory for modern economics, and there is some truth to it. If this is so true, then I believe we should move away from books that focus on general equilibrium, such as MWG, to books that focus on game theory.

Even so, our misunderstanding of ADM will plague our understanding of game theory as well.

This may sound contradictory: I argued in the past that ADM should be revisited, and yet I argue that ADM cannot be the foundation of modern economics done so far. But it is not: we can start from the ADM root and its conventional understanding and formulate a new variant theory.

What is wrong about the current understanding of ADM? There technically is nothing wrong mathematically with the current understanding - the issue is more of a logical problem.

A simple question will answer this: how does ADM treat subjective expectations?

Now you may say that this has already been addressed: because ADM is timeless, equilibrium concepts, such as Radner equilibrium and modern DSGE economics, were developed, addressing shifting expectations and so on.

But even then these equilibrium concepts, while different from ADM, are still, broadly speaking, general equilibrium models, focusing on how "price" changes behaviors of agents, with "somewhat market-clearing" equilibrium forming at the right price vector.
(Monopolistically competitive models, or sometimes called Marshallian models in macroeconomics history parlance, change firm behaviors to not depend directly on the price vector, but then overall they do remain in general equilibrium spirits. A prominent example of course is New Keynesian models, or "policy models.")

And yet they are different from ADM in that they explicitly incorporate how expectations are formed.

And now think of this case: "Agents have different risk aversion, but they share the knowledge of maximum riskless interest/profit rate in the economy."

For this note, I am not going to explain directly why this turns out to be problematic, but then if one carefully examines the issue, one can say that ADM variants in use, especially in macroeconomics but of course not restricted to macro, were blessed with a somewhat ADM foundation and yet cursed with what ADM theory does not actually provide. One of these curses somewhat surprisingly to many is:

 Existence of general equilibrium.

You may say, what? We in econ classes get taught in proofs of existence of competitive equilibrium in DSGE models (or more precisely recursive competitive equilibrium) along with ADM proof of existence of general equilibrium. What are you saying?

Yes, these proofs are mathematically correct. Except, they are not what you are thinking. In fact, in book LS, one of canon books in macro, authors make clear that recursive competitive equilibrium is a partial equilibrium concept, and thus proof of existence only captures the partial equilibrium side.

Thus the only option left inside traditional understanding is: either force expectation formulation to guarantee existence of equilibrium (which really is unnatural) or simply revert to sheer subjective expectation of ADM, which guarantees existence of equilibrium.

Note that the example case involves expectation of something that cannot directly be captured into subjective utility form. ("riskless rate" equilibration/equalization is more of a direct equilibrium condition/constraint than derived inside ADM.)

That completes the memo.

Why I still dislike GNOME 3

A short post on why I still dislike GNOME 3. Adding screenshots later possibly. For now more of a memo.

1. I don't like the so-called minimalist approach. OK, the philosophy that you should be able to concentrate on what you were doing is great. 

 But I just don't like this feeling of having to go to the top-left corner "activities" and have to see application screenshots, and have to press these screenshots to do something. Plus, in some distros, the default setting is that it gets activated just when your mouse pointer in on the top-left corner, which doesn't fare well for people with particular habits (but isn't this a quite common habit?).

2. I sometimes want to concentrate on several tasks done together and wish to minimize and maximize windows very flexibly. And this point partly applies for Unity as well, though Unity is going to be phased out anyway. 

 Of course we can add back the minimize and maximize buttons via some setting tweaks, the question is why this is not default. After all, graphic interface is all about maximizing and minimizing windows and using them as you like, so these buttons are not really worth sacrificing in the name of some minimalist approach. 

3. As in Point 2, I really dislike how the top panel/bar is used for GNOME 3. GNOME 2 and its spiritual continuation MATE use the top bar to have the "menu" interface. And in a way, the traditional GNOME 2 approach still works very well. You want to get through application lists quickly but wish to organize them via folders/directories/groups. The default GNOME 3 actually takes much more space to do this simple task, because it is so icon-based via "activities". Modernity is supposed to make our lives simpler, not just more "fashionable". I really think that Windows and Mac overall do a better job in this part.

 Furthermore, why waste the space of the top panel? After all, all there is for the default GNOME 3 top panel is "activities", followed by time and some background panel applications (usually very few). And time takes some chunks of space in the default setting, which is totally unnecessary. At least for this part, I do think the Unity approach was right, even if history will record Unity as failure. Unity has maximize/minimize icons embedded to the top panel, along with individual application menus also embedded to the top panel, just as in Mac. If you are not going to have "menus" in the top panel, this should be the best approach. Of course the default Unity approach takes the space of the left panel, but for this you can change the Unity panel to be at the bottom, and you are not wasting much space in this approach. 

4. "Desktop". Yes I know that files in the desktop folder can be made to be shown visually in the "desktop." Yet, why is this not the default setting? Minimalist approach? One advantage of course does appear: you are less prone to making mistakes. But I just don't think the advantage is that great compared to sacrifices that must be made. Though I must confess that I myself do not make much use of the visual desktop anyway. (By "desktop" I mean the first thing you see when you enter GNOME environment, just like how you enter a graphical Windows or Mac environment.)

5. And please no, don't say use keyboard shortcuts. There are people that do not prefer keyboard shortcuts because it can mess things up when typing. (activating shortcuts when you are typing that weren't intended) Of course you may say that the same applies for mouse movements, but the former is much likely.

An addendum to the posts on Walrasian equilibrium

An addendum to the posts on Walrasian equilibrium

After writing the posts on Walrasian equilibrium, I realized I should have written paragraphs connecting my first post on Walrasian equilbrium with the second post on it. While responses to No-trade theorem(s) are definitely all sensible, one response should have been much more common: trade occuring on disequilibrium basis.

That is, ex-ante Pareto optimal situation is already hard to occur, and thus every market movement is essentially moving along disequilibrium. This is not saying that people do not realize the ex-ante Pareto optimal situation is required for some of No-trade theorems — people DO. It is more that we do not realize truly embrace starting from disequilibrium condition and ending with possibly disequilibrium.

Does this mean humans are less than “full rationality”? (For now, forget about what it means to be “fully rational”) This is so, in some way. Even with incomplete ex-ante information, agents can engage in a series of exchange contracts so that they do not suffer from disequilibrium problems.

However, and this will connect to some of the upcoming posts, money and modern price-based markets appeared because there are costs to writing contracts and negotiation processes (and I will discuss while-mentioned-often-forgotten costs too in some of the upcoming posts). Furthermore, we can simply assume market structure as being institutional, and assume that agents are doing the best inside that structure.

Of course, as we have seen, disequilibrium inside general equilibrium is troubling, as this does not give us proper allocation, unless we assume that agents can be satisfied with saving money they could have spent to buy something to improve their utility. And excess supply also means “budget/endowment value re-calculation” problem, as some of the assumed budget are not actually available in disequilibrium (but note below). This thus may require a series of re-calculating general equilibrium for agents.

To fix these problems — though eventually one will see, in the upcoming posts and possibly this post also, that there are other possibly more natural options in addition to these — “debts” and “default” arise out naturally, though debts and default do exist in equilibrium models, so some of what appears is really a generic explanation. And this requires modelling expenditure-planning stage, what agents believe as their expected budget, how agents hedge against different possibilities and so on. In expenditure-planning stage, as agents do not have all budgets available to buy things in markets that will open, they will have to take some debts to optimize decisions. In case “price vector” is wrongly set with some agents left with excess supply, agents come to carry “debt” and may have to repay next period (but usually, debt is not re-paid immediately at one time period anyway).

These processes create what Hayek would have called as disequilibrium business cycle (though I am not supporting Austrian business cycle approach here — more of mentioning Hayek because he was one of the first people that discussed disequilibrium-based business cycle approach), with financial sector and real sectors connected tightly.

As said in the previous post, however, some sectors, and here that would be financial sector, are best to be modeled as non-Walrasian. Or this is what many macro models do these days — financial sector thinking “cleverly,” not just relying on price vector for decisions, but relying on prediction models and different information for their operation. Combined with firms also having some thought on how their price vector must be set, this completes a Walrasian disequilibrium picture.

An alternative approach, and this seems justified if one wishes to keep some parts of disequilibrium, while adhering to equilibrium approach in some ways, is to assume that demand drives the economy. This is most represented, at least in mainstream economics, in New Keynesian economics models, even if practitioners may not actively realize the consequences of this. If, for “effective modeling” purposes, firms can be thought of as listening passively for demands, supplying what is demanded, then one can eliminate “excess supply” from the model entirely. One can notice that in “very traditional” sense, this is disequilibrium — however, we can now make use of equilibrium tools.

On Walrasian Arrow-Debreu General Equilibrium and Rational Expectation (2): No-trade theorem and RBC/DSGE models

---

On Walrasian Arrow-Debreu General Equilibrium and Rational Expectation (2): No-trade theorem and RBC/DSGE models

As in the previous story/post, when one says Walrasian analysis, it refers to Arrow-Debreu general equilibrium analysis, or shortening it, general equilibrium analysis. When I talk about DSGE models, I will refer to them separately — because while DSGE models are in some way rooted in general equilibrium anlaysis, they are not entirely rooted in general equilibrium analysis. This post intends to think about this.

It is well-known that if trade is ex-ante Pareto optimal, and human rationality and information structure is common knowledge (people have incomplete information of the world), then even if information changes, there cannot be any trade going on. This is what one calls as No-trade theorem, and for now I will not discuss more results on No-trade theorem. Suppose agents were previously in “ex-ante” general equilibrium. If information then changes, no trade will occur to change the allocation by No-trade theorem. (price vector changes, if the Walrasian market is to re-open after new information is revealed to some agents, however.)

It does seem that No-trade theorem restricts much about what general equilibrium markets. Once Arrow-Debreu trade occurs at time t=0, no trade is required, nor can occur. However, there are ways to get around this while maintaining the reasonable assumptions imposed by No-trade theorem, and of course those said in this post are not all of them.

If some to-be-realized states cannot be learned by agents, and agents cannot even know their existence is possible, then even after a trade already occurs, a new trade may occur. An obvious example is introduction of new products totally unanticipated by everyone.

Also, if agents can believe that they are not in state x even though they are in state x, then No-trade theorem can break down. This one will not be discussed.

Let me return to the question of some states that possibility of existence is not known. Essentially DSGE models re-open Walrasian markets at every period — reflecting the fact that there exists new products (even though models often have homogeneous goods, or representative goods). Since there are infinite periods in many DSGE models, this means that Walrasian markets are opened infinite times, while usual Arrow-Debreu general equilibrium analysis has finite number of goods. Here change in expectations has important consequences for each period, unlike in Arrow-Debreu general equilibrium, which lacks “business cycle.” It is obvious why DSGE models are used for business cycle analysis in this way. Essentially, modern DSGE models, that mostly can be said to be RBC variants, are incomplete market models.

However, one must be reminded of Maskin-Tirole critique (Maskin and Tirole (1999)), which roughly states that parties can write “contracts,” here including state-contigent trade promises, such that they know states even though they do not. Thus the above possibility of going around No-trade does pale somewhat. This, however, does not mean that the above possibility is wrong — rather, it is that necessity of deviations of incomplete market results from complete market results does not necessarily follow from unanticipated states/contingencies. We may still accept incomplete market as a feature of our reality and do analysis, at least in macro terms, and forget about theories of firms that investigate how firms arise.

---

Addendum: I may discuss Maskin-Tirole critique and some industrial organization/new institutional economics topics in a separate post.