Unintended Consequences and the Social Sciences

18 April 2007 at 3:57 pm Nicolai Foss 13 comments

| Nicolai Foss |

According to a prominent tradition in economics and classical liberal thought, the social sciences (particularly economics) are primarily concerned with explaining unintended phenomena, whether more temporary outcomes, such as market phenomena, or more permanent ones, such as institutions, in terms of the intentional actions of multiple interacting agents. In contrast, the social sciences are not really taken up with explaining individual action per se.

This is an understanding, or perhaps even doctrine, that is perhaps most famously associated with Hayek, but it has also been echoed by Ludwig Lachmann (among Hayek’s contemporaries) and by many modern Austrians, as well as by philosophers, notably Karl Popper, Michael Polanyi, and Edna Ullman-Margalitt, and (non-Austrian) economists such as Andrew Schotter (in his 1981 book on institutions). (Of course, the notion of unintended consequences itself is by no means unique to classical liberal economists but can be found in the thought of most major thinkers on social science and political philosophy).

However, there are several problems with the doctrine that the social sciences are mainly about unintended consequences. Here are two that seem to have not been previously noticed:

First, to the extent that aggregate outcomes are designed by a conscious will — such as, in the extreme case, the mythical omniscient and benevolent planner that still plagues many economic models — the unintended-consequences view rule them outside social science because the outcomes are direct consequences of the intentional actions of a single agent. A presumably unintended consequence of the unintended-consequences approach is that mechanism design and agency theory must lie outside of economics. This is patently absurd, and luckily the revealed preferences of the Swedish gentlemen in the Nobel Committee have shown they disagree (i.e., Nobels to Mirlees and Stiglitz — and Kantorovic and Koopmans).

Second, the notion of unintended consequences makes appeal to ignorance (particularly in the case of Hayek’s thought) — “unintended” usually implies “unforeseen,” something that did not enter the deliberation of the interacting agents at all.

However, this seems descriptively rather unrealistic. Many of us have been exposed to examples of presumed undesigned institutions, ranging from paths through a wood to morality itself. Take the mundane example of a path that emerges through a wood. While the path emerges, does it emerge in an unforeseen manner? Doesn’t somebody who walks along the emerging path want it eventually to become a path (because of the added comfort of walking along a path)? No ignorance/unforeseen consequences need be involved in this case. In fact, the emergence of the path is not really unintended; on the contrary. I suspect that this example can be generalized (e.g., to the Mengerian account of the emergence of money), and what is really at stake is in many cases not ignorance at all, but rather the notion of externality. Thus, what we call “unintended consequences” may often better be called “market failures” (I know Tyler Cowen has made this point somewhere, but I have forgotten where).

Entry filed under: - Foss -, Austrian Economics, Classical Liberalism.

The As-Is Journal Review Process New Paper by Hart and Moore

13 Comments Add your own

1. Peter Klein | 18 April 2007 at 4:20 pm

To add to Nicolai’s first point:

Last year Dan Klein and Pete Boettke debated retiring the term “Austrian”:

http://www.gmu.edu/departments/economics/pboettke/workshop/Spring_06/Plea_to_Retire_Austrian.doc

http://www.gmu.edu/departments/economics/pboettke/workshop/Spring_06/Boettke_Austrian.doc

Dan proposed replacing the label “Austrian economics” with “spontaneous order economics” or “sponecon.” In an email to Dan I noted the following:

“I’m sympathetic to your arguments against the ‘Austrian’ label and would be happy to abandon it for something else. But I’m not convinced that ‘sponecon’ does the trick. You note at the end of your piece that ‘sponecon’ appears to omit the analysis of interventionism and politics; but what about production theory, business strategy, and the economics of organization? That body of literature — to which Austrians are increasingly contributing — deals primarily with the economics of authority and delegation, not spontaneous order (‘market-based management’ notwithstanding). Overall, I’m not sure that ‘sponecon’ characterizes the *essence* of Austrian economics.”
2. Gary Shiu | 18 April 2007 at 9:02 pm

With regard to Nicolai’s first point:

I would imagine those who buy into the unintended approach would argue that agency theory or mechanism design theory might still be relevant at the firm level though. So it seems the issue hinges upon the level of analysis (the whole economy or macro if you will or at the micro-units that are interacting to produce the aggregate outcome)

With regard to Nicolai’s emergence of path story:

Framing the path story in unintended consequence terms, it would go something like this:

A guy is walking with his dog along a well-treaded path. The dog suddenly runs off the well-treaded path into the woods. The guy chases after the dog into the woods. A new path emerges not for the sake of creating a new path on the part of the guy but as a byproduct (unintended consequence if you will) of his attempt to find his dog. Other walkers become aware of this new path and find it convenient to use. Usage of the new path becomes the norm for all walkers.
3. Daniel Klein | 18 April 2007 at 9:44 pm

RE Peter’s remarks, I’ve reconsidered “spontaneous order economics” as a name. Just too long, and no good abbreviation.

But I still think that the character is true. Graduating from character to identity is something I discuss in the following Econlib essay:

http://www.econlib.org/LIBRARY/Columns/y2007/Kleinidentity.html
4. Nicolai Foss | 19 April 2007 at 12:45 am

Gary, I am not sure your recasting of the path story qualifies as a real spontanaeous order/invisible hand explanation. In what is perhaps still the best treatment of the finer points of invisible hand explanation, Ullman-Margalitt (1977, Synthese) stipulates that the explanandum of such an explanation must arise as the unintended consequence of the interaction of multiple intentional actors. In your story, the path is basically there once the the dog and the guy has created it. It is not created as a result of the interaction among multiple intentional individuals per se (we can exclude the dog from the set of agents we consider, I suppose).
More generally, I don’t think the point of invisible hand explanations is to stress the consequences of a single persons random actions , unintended by himself only; rather, the point is to stress how nobody in a large social system may intend and foresee an outcome, yet it happens as a result of their actions. This seems different from your example.
5. Gary Shiu | 19 April 2007 at 3:07 am

Nicolai:

Thanks for the reply.

But does your argument also applies to your own example is equal force. The path story of yours has nothing to do with unintended consequence type of story precisely because there is only that “somebody” in the story who intends to tread a more covenient path than that already in existence?

I guess one way to state your position is that even if a game theory type situation where multiple equlibria exist, the eventual settlement on a particular outcome is still not qualified as an unintended outcome because players of that game are aware of the multiple equilbria nature of the game before its start, does this fit in with you are thinking in the above post?
6. Nicolai Foss | 19 April 2007 at 3:16 am

Gary: The point of my example with the path was to suggest that many of the examples that we use to demonstrate unintended consequences may downplay intention, rationality etc. _too much_, Thus, those agents who makes the path emerge as a path may very well all intend it to become a path. They create positive externalities by their actions.

Your game theory interpretation seems right to me, for the reason that invisible hand explanation seems fundamentally associated with ignorance. Thus, the eventual equilibrium is unforeseen by the players; they are fundamentally ignorant about it. Standard game theory cannot allow for this possibility.

More generally, I think we should focus research on invisible hand explanations by discussing the boundary conditions of such explanations in terms of how much we can allow agents to know about the outcome of their interaction. Presumably, unless the interaction in question is entirely parametric, knowledge of the outcome will generate strategic behaviors, which makes the outcome foreseen and partly willed. So, how stupid need agents to be for invisible hand explanations to be valid?
7. twofish | 19 April 2007 at 10:44 am

I’m interested in this topic from the point of view of trying to figure out what caused great social disasters and how to prevent them.

I think the most interesting “unintended consequence” situations are those in which the consequence is *contrary* to the intentions of the people who devise the policy. Iraq is an example. Communist societies are another.

I think what is important is not “ignorance” but “meta-ignorance” or what the Greeks called hubris. The common theme with great social disasters was not that the planners did not know, but rather that the planners did not know that they did not know. If you start with the assumption that there are large parts of the problem you don’t understand, you are able to make adaptive corrections. If you don’t, you can’t.

Part of the reason why China succeeded whereas Russia failed, was not because of specific policies, but rather because of philosophical framework. Chinese economic reform started from the premise that no one really knew what the correct policies where, and so massive experimentation was necessary. The big problem with Russia was that since Communism had supposedly created a perfect socialist society, there was no way of admitting problems and allowing experimentation, without bringing the whole infrastructure down.
8. Otto BP | 19 April 2007 at 3:38 pm

Dear Nicolai,

Good points. Two questions, though.

1) Why should ‘unintended’ necessarily read ‘unforseen’? Is it really important to the argument? Certainly, by virtue of your externality point, agents must to some extent ignore at least some aspects of the consequences following an act (such as making a path in a wood).
2) Your externality point is well taken. But ‘external effects’ don’t necessarily translate into ‘market failures’. In order for market failures to occur, external effects must be present at the margin. It is easy to think of examples of unintended infra-marginal consequences. Just think of any regular market. In Adam Smith’s story about the market, the consumer surplus is an unintended consequence of profit-seeking by the baker, the butcher et al (just as profits are unintended consequences of consumer demand). Both parties infer external effects on each other. Except at the margin, where there are no profits or consumer surplus.

Otto
9. Nicolai Foss | 19 April 2007 at 4:08 pm

Otto, Excellent comments! In response:

1) Logically you are of course right: “Unintended” does not necessitate “unforeseen”. However, Ullman-Margalitt clearly associates the two, and I think that most thinkers on unintended consequences have made the association. It is obviously central to Hayek’s thought .

2) Right! Externalities do not automatically translate into market failures.

3) As to your examples of infra-marginal surpluses, can we really say that the consumers or producers surpluses are “unintended”? Presumably firms very much intend profitability. Some measure of market power (if you buy the traditional story) allows you to not only intend, but realize to a certain extent profit. By means of search and/or bargaining, consumers can very intentionally influence their surpluses. Etc. They may be frustrated in their attempts, but does this make the final outcome “unintended”?
10. Otto BP | 20 April 2007 at 3:53 pm

Dear Nicolai,

3) You are right, of course, that profits are very much an intended consequence of firms seeking profits, as are consumer surpluses an intended consequence of consumer behavior.

The point I was trying to make, however, is that consumer surpluses are also allways an external effect (or unintended consequence) of firms seeking profit (and vice versa, in the case of profits). (I had this quote from Smith in mind: “It is not from the benevolence of the butcher, the brewer, or the baker, that we expect our dinner, but from their regard to their own interest.”). In fact, any economic relation between two or more agents could be said to always involve external effects, but in most cases they would be infra-marginal only.
11. Sudha Shenoy | 23 April 2007 at 8:01 am

Hayek distinguishes between ‘taxis’ & ‘cosmos’.

A ‘taxis’ is an organisation like a club or a company, with definite objectives. Eg, a club may be ‘social’ or for playing golf, or chess, viewing foreign films, reading books, whatever. A club’s objectives may change in course of time, as members’ views change. Similarly, a company produces one or more goods &/or services. And these too may change in due course.

A ‘cosmos’ is eg; a language, a market order, the common law, moral rules, customs, etc. It ‘grows’ through the actions of individuals; is only found therte; is an instrument for their several purposes. Language exists only in its speakers (so to say); developes through their use of it; is an instrument for their several concrete communications.

Hayek developed further the insights of Menger. Mises too distinguished between ‘organism’ & ‘organisation’ (in Socialism.) The recognition of two kinds of orders — abstract & concrete — goes back to the early- to mid-seventeenth century debates between the common lawyers & Hobbes.
12. shay | 15 August 2007 at 2:50 am

The notion of unforeseen and unintended does not refer to ignorance, rather it just points out man’s limitation. We cannot comprehend the future 100%. That is what Hayek pointed out.
13. Nicolai Foss | 15 August 2007 at 8:19 am

Shay,
No. You are wrong. In several ways:
1. Hayek’s argument is about what may be called “currently existing knowledge,” rather than about the future. The Misesian tradition in contrast places more emphasis on the future.
2. Even if Hayek had indeed primarily stressed the future, why wouldn’t that be “ignorance”? Of course we are ignorant about the future, since it hasn’t been created yet.