Give Me Placebos Or Give Me Death:

Neuroskeptic has the story, and it appears not to be simply because healthy people are also more responsible, they controlled for measures of health.

 

(Via Cheap Talk)

Open science: why is it so hard?:

[Snip…]

Thus, a much more significant vision is Nielsen’s open science. Michael Nielsen is arguing for a culture shift in science: from a science obsessed with individual performance (and publications) to a science culture resembling more that of open source software or wikipedia.

I fear however that despite all the (well deserved) press that Michael Nielsen’s latest book has been getting, too few people understand the importance of this shift. It is not about becoming hippies. It is not a socialist utopia. On the contrary, the system we have right now is akin to an highly regulated industry. All power is in the hands of the government and a few large organizations (universities, publishers) working in tandem. The barrier to entry is maintained artificially high. Open science is really about creating “open markets” with freer exchanges. It has the potential to boost our collective productivity by orders of magnitude through the removal of unneeded friction.

[Snip…]

And we finally get a hint at why it is so hard it is to open up science: the business of science has become intertwined with businesses like the publishing business. ACM has to speak both as an association of computing professionals, and as a publishing house.

What should be a critical support service, the publication of results, ends up driving much of our culture. The journals become the science. The medium becomes the message.

In effect, we have too much organizational scarring tissue in science. It could be that we need to reboot the system. As a starting point, we should collectively recognize the problem. Repeat after me: scholarship is not a publishing business.

Further reading:

• Is Open Access publishing the solution? Really?
  
• ACM’s role in public policy
  
• Time for scholarly publishers to disavow the AAP

Update:

The ACM charges the authors of any conference for the publication of proceedings. However, they do not require payment for publishing in their journals: instead they request page charges.

How to revise research papers after receiving harsh reviews:

Whether you submit your work scientific journal or just post it on a blog, you can expect to receive harsh criticism from time to time. Sometimes you are facing arrogant or ignorant readers. Other times, your work is genuinely flawed. My own work is frequently flawed, as you know if you read this blog.

Over time, I have learned that even if the reviewer is wrong, spending time to careful respond can be tremendously useful. If you are 100% correct, then you get to build up your confidence and can later answer similar criticism hastily. Very often, however, you did not do everything perfectly. Maybe your arguments and data are correct, but you might have presented them better.

There are specific strategies to deal with harsh reviews:

(Snipped. Excellent practical advice is here, but I really think you should go read the original post!)

(Via Daniel Lemire's blog)

ADDENDUM: hat tip to Ole Rogeberg on Google+.

If you wonder why I have not gone over to my university's online library site to read the review, it's because I am utterly disgusted by the Williamson quotations and do not want to waste my time reading his review. Macroeconomics is supposed to be a science, not a part of analytic philosophy or logic; "economists" with such enormous blindfolds as Williamson have too much sway in the discipline and have corrupted its very core.

Like Tjalling Koopmans said in response to Milton Friedman's methodological emanations, every assumption you make in building a theory is automatically, and rather obviously, also a prediction of the theory. Saying "I like the assumption of rationality, I will always make it, and it's my mode so you can't tell me not to play with it to my heart's content" is odious nonsense. When the efficient market hypothesis or DSGE is part of a model that produces predictions that keep being smacked by the data, insisting that these assumptions (and since when is DSGE an assumption, rather than a whole modeling technique?) can't be wrong is tantamount to saying that your theory can't be useful and in fact is eminently ignorable. If you peddle it to the world and your students as science, you are at the minimum corrupting the notion of science itself.

OOP is a style of writing computer programs, with some languages, such as Java, heavily supporting it (Java enforces it, in fact) and other languages, such as Python, supporting it strongly. Coming to OOP from economics was conceptually easy. Now I am thinking that OOP has something to tell us regarding how we teach economics or how we present it to the wider world. This post is my first written rumination on this topic.

First, let me define an object in the OOP sense. I need to give you some background first. I expect you are aware, even if you never wrote any computer code, that such code is a string of data, ultimately encoded in binary notation, that tells the computer to do certain things with part of this data. To take a simple example, if you want the computer to choose randomly between the names of participants in a lottery, you have to include the names in your program (data) and write commands that result in one of these names being chosen randomly and this choice being communicated to the user. This little program has data (it "knows" some things) and it has ways of acting on these data (it "knows" what to do with the things it knows).

In any reasonably complicated programming problem, it helps us as programmers to compartmentalize the code. We make a chunk of code to perform task A, another to perform task B, and so on, and finally we write code to coordinate these chunks as they go along merrily doing their thing. Each one of these code chunks has some data it knows and some things it can do with the data it knows. (Computer scientists, I know I am simplifying. I only want to convey the basics of OOP here.) An object in OOP language is a chunk of code that has some data and some things it can do (usually called "methods", but you can think of them as commands specific to this chunk). The programming language in which you are specifying these objects provides ways for objects to communicate with each other, passing data around, and to ask each other to perform one of its methods (execute one of the commands it has).

How is this relevant to economics? You are probably already chomping at the bit to answer, but here is my take. When we set up any model in economics, at least if our model is "microfounded" as we say, we have some agents who know certain things and do certain things. Suppose, for example, that you want to make a computer model of an exchange economy. You need a number of individuals, each with an endowment and a preference relation over the commodity space. These individuals need to be able to perceive prices, decide which net trades are feasible given a particular price vector, and, finally, they need to be able to propose and execute trades with each other. (OK, maybe you wanted a Walrasian auctioneer thrown in as well? I leave it as an exercise then to determine what data and which methods the auctioneer object in the code will need.)

I cannot think of a mainstream economics model that cannot be conceptualized in these terms. Indeed, computational economics has flourished in the last several years and you can find plenty of examples of what I am talking about by visiting, for example, the amazingly comprehensive website that Leigh Tesfatsion has set up. (This kind of "microfounded" computational economics goes by the name agent-based computational economics.)

So why don't we teach our students using the concept that each economic agent in our models is an object (in the OOP sense) that interacts with others based on rules set up by the institutional infrastructure of the economy and physical feasibility conditions?

I can think of two answers right now, one deeper than the other. I will examine each one and argue that they do not convince me, starting with the shallower one.

The shallow answer is that our students will find it hard to understand the OOP language of objects and their interactions. I am much more likely to agree that our students in introductory courses will not understand supply and demand graphs or the simple linear equations we try to use to overcome their math anxieties. But almost all students in introductory courses will have played computer games. For them, all you have to do to introduce OOP objects is to refer to an avatar in an online role-playing game or a tile in Tetris. Stepping from these examples to explaining economic objects in code does not in fact impose the need for a computer language at all. You can use pseudo-code and the ideas stand; any time your students feel shaky, just bring in another computer-game inspired example.

The deeper answer is that while encapsulating an individual in an exchange economy in the language of OOP objects is easy, if we start thinking in the algorithmic terms this mental shift suggests, things like arriving at a Walrasian equilibrium become hard problems. There might even be a student in your class who knows enough computer science and will tell you to your face that our cavalier approach of taking shortcuts like calculating an equilibrium with Lagrange multiplier techniques and setting supply equal to demand is a rotten approach, as it hides the remarkably difficult problem of arriving at an equilibrium.

So how can I deal with this answer? My point is that we should adopt the OOP viewpoint precisely so as to force our students or readers to confront the fact that reaching an equilibrium in an economic model is much trickier a proposition than the typical paper in Econometrica or JET lets on. As economists, we have internalized the mental shortcuts that make us jump to equilibria in fairly complicated models and then analyze the properties of these equilibria. But there is much to be learned by confronting the need to specify exactly how economic agents interact in time, each with its own data an abilities to perform actions such as buying and selling, manage to get to an equilibrium (if they do). Do we really want a Walrasian auctioneer who gropes around in price space to find an equilibrium? What if the economy happens not to have a stable equilibrium (the question can arise whether we are looking at a market model or any other kind of agent-based model). These are not idle concerns; they show clearly some of the limitations of economic theorizing and to ignore them is intellectual arrogance at best, dishonesty at worst.

There is one final point in my mind about this, which I will leave to be developed in a future post. Thinking in terms of institutional infrastructure can be considered the overall code of our economic computer program. This can encompass the ways information gets passed around from agent to agent as the economy operates as well as the outcome function that determines what allocations occur and when as the agents take various actions. The more precise we are in specifying these the better, just as much as the more careful we are to specify computable ways to reach an equilibrium (see the previous paragraph) the better. As this post already exceeds 1250 words, however, this final point will have to be explored later.

In the end, there is a deep contradiction in Hayek’s thought. His great insight is that individual human beings muddle along, making progress by planning, experimenting, trying, failing and trying again. They never have as much clarity about the future as they think they do. But Hayek somehow knows with great certainty that when governments, as opposed to individuals, engage in a similar process of innovation and discovery, they will fail. He insists that the dividing line between state and society must be drawn according to a strict abstract principle rather than through empirical adaptation. In so doing, he proves himself to be far more of a hubristic Cartesian than a true Hayekian.

UPDATE (2011-05-09): Although the commenter to this post did not respond to my request to show me who has contradicted Fukuyama online, I did find a whole lot of comments via Easterly's blog. This one, from beyond the grave, is most entertaining. Do read the comments on Easterly's blog, too. At the end of the day, Fukuyama's review, at least the part I quote above, is certainly true about people like Beck and may not be true for Hayek. I stand by my observation in the comment below that Hayek should have looked for the mathematical tools needed to formalize his seminal ideas. (I have almost the same complaint about Keynes's General Theory, although Keynes was reasonably capable in mathematical modeling. Why did he not try it?)