Why Fully Determined Systems Aren’t Necessarily Boring

Last year, we had a professor at the University of Wisconsin named Daniel Bromley give a lecture on institutionalism. Bromley, out of the old institutionalist style of Commons and Veblen, was very critical of Douglass North’s work, and the reason for that, he said, is that North is seeking to endogenously explain institutions. If he is successful at doing so, argued Bromley, he will have effectively turned the world into a deterministic system. My goal is to critique this theory and consider it lacking in light of the nature of a spotaneous order.

To begin with, let’s discuss institutions. An institution, simply put, is a rule, be it formal or informal. Shotgun is a very informal rule, whereas the American constitution is made up primarily of formal rules. According to North, these institutions develop over time, and one of the things that he has attempted to do is find a way to show that the institutions which prevail in a particular society were created because of previous institutions. Thus, the United States has the rules that it does because of its history, and the only way to have the same set of institutions as the United States would be to have the same history. This would cause somebody like North to have grave doubts concerning the efficacy of attempting to import institutions and have them stick (such as the example of Latin American countries copying the American constituiton—yeah, they have the same words, but the institutions, particularly of the informal variety, are different). But more improtantly, if there is an entirely endogenous explanation for the rise of particular institutions, you should be able to predict what will occur over time, given this complete theory, the present institutions, and any additional circumstances which need to be factored in to the system. According to Bromley, this would create a deterministic system, and this would be intensely boring, as you know exactly what would happen. But I’m not so sure that this is actually true.

Let us assume that my description of Bromley’s description of North’s theory is completely accurate as far as reality is concerned: there is a fully endogenous explanation of institutions, and furthermore, we have all of the information available to plug things in to our model of institutional “general equilibrium”* (which is a stretch of the imagination, but that’s another issue that I could discuss later). So we know exactly which institutions will exist at any point in time, and as far as Bromley is concerned, that’s the end of things.

But for me, that’s not the end of things. We’ve described the order of rules, but we have left out the order of actions. Let us hearken back to Hayek’s notion of the spontaneous order, using an example from nature. We have a good idea of the rules of physics, and know, for example, how to form crystals. A crystal is an example of a spontaneous order, as there is no force which willfully acts to create the order of molecules in the crystal, and yet this result occurs. Furthermore, it is not a haphazard result—every time you provide the same circumstances, you will get a crystal to form. The difference, however, is in the placement of individual molecules. Even though we have determined that a crystal will form, and perhaps we even have an idea of its shape, there is no way to determine exactly which molecules will be in which positions using the rules of physics. On the level of actions—what individual molecules end up doing—there is a level of randomness even though the order of rules is fully determined.

To bring this metaphor up to the level of individuals interacting in a society, even if all of the rules are fully determined, it still has little to say about what any individual in a society will do. The set of formal and informal rules in place influence greatly the scope of human action and even assist in forming preferences, but when you are talking about the realm of human action, even a highly-ordered society must leave a large amount of scope for human action given the complexity of everyday life. For example, even if Douglass North or Daniel Bromley could tell me the exact nature of institutions as they will exist in fifteen years in the United States, neither could tell me what I will be doing in fifteen years, and neither would have much luck determining the particulars of my behavior. They might circumscribe boundaries to my behavior, but these would necessarily be vague and general, as the institutions which exist do not allow for any more. So just like how a chemist could give some rough information about the location of a molecule as it goes from existing freely to becoming a part of a crystal, an institutional economist could give some rough information about how individuals behave in a society. But given a set of rules, you cannot determine the actions which will result from the rules in a non-trivial, complex order.

So on the level of the chemist/economist, he may see the order as fully determined on an abstract level, but on the level of the molecule/individual, there is still a great deal of randomness involved. And even at the level of the chemist/economist, “fully determined” means less than you think: it is a full determination of the rules, but certainly not of the actions. Because of this, I would suggest to Professor Bromley that he should not lose any sleep over such an idea, for even if institutions were entirely determined, it would not really make life boring.

Incidentally, I doubt that North or anybody following him will come up with such a deterministic theory, as there are too many exogenous factors involved (such as weather, migration patterns, wars, natural disasters, unforeseen technological developments, and so on—in other words, Kirznerian discovery plays a role here, too). To the extent that we might have endogenous theories of institutional development, they will be very vague and will not actually be able to be used for legitimate prediction of future institutions.  In fact, they would be similar to theories regarding the development of spontaneous orders:  predicative only on the highest levels, but highly undetermined at the lower levels and almost entirely undetermined on the level of individual actions. But that, too, is something that I could save for another post, another day…

* – Please excuse the term, but it’s the first thing that came to mind…


The Silliness Of The “Marketplace Of Ideas” Concept

Oliver Wendell Holmes famously defended free speech by alluding to the “marketplace of ideas.”  He did not actually use these terms, but Keyishan v. Board of Regents did put it into the written judicial code.  The basic concept of the marketplace of ideas is that ideas are competing against one another, so the good ideas will prevail over the bad ideas, just like how more efficient production processes, more innovative individuals, and more appropriate products will win out against those worse processes/products.  The downfall of this concept is that the realm of ideas really _is_ like a marketplace, but people who use this term usually don’t have much experience with what a marketplace really does.  In actuality, the functioning of a marketplace often leads to precisely the opposite of what “marketplace of ideas” proponents believe.

To understand a marketplace and its functioning, you have to understand the people who make it up.  Individuals in a marketplace can be abstracted to have a few qualities:
– Preferences
– Constraints (most typically, the budget constraint:  I can only spend as much as I make)
– Information regarding prices, available quantities, and characteristics/properties of goods
– Capacity to acquire goods (for example, I can’t buy a pre-fab house which is bigger than the lot that I own or put in a pool in a neighborhood which bans pools)

All of these are subjective.  Preferences are highly subjective.  Constraints are subjective—Bill Gates has a different constraint than, say, Dan or Tony.  Information is subjective, as some people are more knowledgable about particular pieces of information than others.  For example, ask a gaming afficianado about how much a GeForce 8800 GTS card costs and he’ll either tell you on the spot or go to Pricewatch and find out that  they’re running at around $290.  Ask a person who refers to a computer as “the hard drive” how much one of those cards cost and they’ll look at you funny.  Finally, the capacity to acquire goods is different for different people as well.  This could be fit right into constraints, although economists generally ignore this one by trying to avoid any reasonable use of time (so basically, you select your bundle of goods all at once and there are no lock-in effects).

So if preferences, constraints, information, and capacity are all different, we can surmise that people will want different things.  Again, this is hardly novel—a computer geek will be more concerned with having up-to-date computer hardware and perhaps reading computer-related magazines, whereas a car buff will spend more of his time and money on automobiles, automotive magazines, and the like.  So the purchases will be different.  But on top of that, there will be heterogeneity _within_ products as well as a range of products.  With video cards, take a look at the Pricewatch link and see how many cards there are.  There are cards which are primarily for hardcore gamers as well as cards for business computers, notebooks, and for people who want to watch TV on a computer.

Another example is roofs.  If I look outside my window I can see different types of roofs.  Some of them have tar-and-gravel shingles, which are very common in the US.  Some of them have thick ceramic plates.  The house I am living in has what looks to be stone.  There is a tin roof for a shed, and a concrete one for a flat-roofed building.  Finally, a church not too far from here has a different type of stone roof.  Just in this small area, there are 6 different kinds of exterior roofing materials.  There is a reason for many of these that even I, hardly a roofing expert, can gather:  the concrete roof is flat to act as a patio; the tin roof is for a shed because tin roofs are very inexpensive and you don’t live in the shed, so you don’t care about the pinging noise whenever it rains; the tar-and-gravel roof is self-sealing and lasts a while, though the building does get hot; the ceramic plates keep water out like tar-and-gravel but don’t get as hot; the stone acts similarly; and the church has its own roof because it was built before modern roofing practices came into play.

So in our world, even similarly-designed houses in the same area can have strikingly different roofs and people use quite different computers based on their needs.  Our world, to steal a concept from Ludwig Lachmann, is a world of two types of competition:  first, we have a new product come on to the market and you have different firms providing similar versions of this product; but once the product becomes a bit more developed, you can start to see different firms trying to differentiate their versions.  As an example, Coca-Cola hit the market in the late 19th century.  After it entered and started to gain some success, other firms entered the market and billed themselves as similar to Coca-Cola.  Even in the 1950s, for example, Pepsi was considered little more than a generic Coke, and the firm didn’t do much to challenge this idea.  But after a point in time, as the market for soft drinks solidified, you started to see product differentiation.  Pepsi, RC Cola, and several other versions of cola began to advertise in such a way as to distinguish them from Coke.  Pepsi turned out to be quite successful at appealing to younger purchasers, particularly by the 1980s, and they were able to segment the market.  This goes to show that the basic neoclassical concept of competition—in which firms produce homogeneous products—is not completely wrong, but is quite incomplete*.  In the world as it exists, markets lead us to heterogeneous goods.

So, what does this have to do with the “marketplace of ideas”?  Well, people have different preferences, constraints, and information here as well.  In this case, the constraint is our capacity for understanding ideas.  Information is just the same as before—you have to know where to look to learn ideas, and what you know might be wrong.  Finally, individuals have preferences.  The easiest example of this is political preferences—people who tend to be right of center are likely to want to read more people who are right of center and vice versa.  People who don’t have well-formed political beliefs, meanwhile, are not likely to want to pay attention to politics at all.

So given that, at least from this perspective, there seems to be a “marketplace of ideas” in the sense that ideas appear to form a market just like other markets, what is my point?  Well, the point is that you have to look at both kinds of competition in markets:  product innovation and product differentiation. The concept that most people who use the term “marketplace of ideas” want to explain is that they like to see the realm of ideas as one in which competition destroys the bad ideas, similar to how markets destroy inefficient production processes.  But this misses the entire strand of competition dedicated to product differentiation, taking advantage of different preferences.  As I said, people are more comfortable reading things with which they agree, and entrepreneurs in the market for ideas understand this, so they cater to certain subsets of the population.  Thus, far from destroying the inefficient ideas, marketplaces can allow for niches.  For example, Marxism was an ideology based on a theory which was shown to be patently false almost 140 years ago, and yet there are still Marxists today.  The marketplace of ideas works just like any other market:  entrepreneurs of ideas (Marx, for example) provide positive value to some group of individuals whose preferences lead them to desire something along these lines.

So, far from being an engine of creative destruction and homogeneity, markets in ideas—just like markets for other products—develop into product differentiation based on subjective preferences.  And for this reason, the apt metaphor does not actually mean what you think it means.
* – I should note that there are more advanced models based on oligopoly and monopolistic competition which handle the idea of product differentiation, but a solution in such a market is still considered “sub-optimal” because the only optimal solution is in the first type of compitition, or in special industries such as (perhaps) agriculture.

Installing Mono—What Gives?

Okay, so I went to install Mono. I saw that there was an entry for Fedora Core 5, but none for FC6, so I kept looking. Then I saw the simple installation binary and figured hey, why not be lazy. Well, I installed the binary just fine and went to run monodevelop, wanting to see what it looks like. Err…it looks like a crash. After doing a fair amount of debugging, I eventually found out that the monodevelop which ships with the installation program requires gtk-sharp2 version 2.8 or above. The version of gtk-sharp which ships is 2.4. Why? I don’t understand this, and it seems as though it would deter anybody who would really need the binary installer because they won’t have too much of a clue what’s going on.

Perhaps I was doing something wrong, but I don’t think so. At any rate, I went and installed it using the Fedora Core 5 RPMs and now everything is copacetic, but it’s more than a little weird that one of the more important pieces of the Mono set—the IDE—would not work out of the box due to library issues…

Well, actually, not everything is copacetic.  For some reason, gnome-sharp is installed and in the GAC.  If I run gacutil -l | grep -i gnome-sharp, it shows up.  But things don’t show up in monodevelop.  Huh?  I think I need to go onto IRC and figure this out.  If I get a solution, I’ll update this post.

Rebuilding The RPM Database

Using Fedora Core, it seems that the RPM database goes and corrupts itself sometimes. I don’t know why, but at least there’s a solution to it. First, log in as root (or use sudo, which actually is a better idea). Then run the following commands:

rm /var/lib/rpm/__db*
rpm --rebuilddb
yum clean all

If you want to watch everything, change the middle command to rpm -vv –rebuilddb. Watch out, though, as this will make things slower, as I/O calls are pretty slow, even when they’re just going to the screen.

What the –? OH MY G_D, IT’S A GHOST CAR!!!!

I shall be relocating to a new lair tomorrow, so I may be without the internets for a couple of days (maybe more!) If you’re beating down my door, waiting for cake, you shall get your cake, my friends. I shall return soon, though, and Kevin will probably be peppering the site with posts (and Dan might put up his defense of Christianity; I’ve seen a sneak preview, and it’s everything you’ve come to expect from him!)