Jako zadanie domowe proponuję sprawdzenie, jaką strategię powinien wybrać przedsiębiorca produkujący skarpetki w poniższej grze z Naturą. Rynek spełnia warunki definicyjne Natury: podejmuje decyzje, które wpływają na wynik Producenta, a jednocześnie jego reakcja na wybory Przedsiębiorcy jest tak znikoma, że dla naszych potrzeb możemy ją pominąć. Proszę sprawdzić, jak powinien zachować się przedsiębiorca przyjmując za kryterium wyboru:

1. Kryterium Laplace’a

2. Kryterium Walda

3. Kryterium Hurwicza (współczynnik optymizmu α=0,75)

4. Kryterium Savage’a.

Zastanówcie się też, które kryterium jest najwłaściwsze z punktu widzenia Producenta skarpetek grającego w tę grę z Rynkiem. Powodzenia!

Die an dieser Stelle im vorigen Beitrag dargestellten Entscheidungsverfahren bei Unsicherheit haben einen wesentlichen Nachteil: sie sind “ganz” optimistisch oder “ganz” pessimistisch. Das macht Minimax und Maximax aber gleichermaßen unrealistisch, denn die wirtschaftliche Umwelt erfordert oft, sich nach Maßgabe der Situation etwas optimistischer oder etwas pessimistischer zu verhalten. Genau das leistet die auch als “Pessimismus-Optimismus-Regel” bekannte Hurwicz-Methode.

Mehr…

¿Eres optimista o pesimista? En economía de la empresa estas dos opciones son las que se plantean para resolver ciertos problemas de decisión. Dadas diversas alternativas de decisión empresarial, junto con sus resultados esperados en función de sucesos ajenos a la empresa, el criterio “optimista”, el que se supone que utilizaría un hipotético gerente con un 100% de optimismo, se centraría en los mejores resultados, y escogería aquélla alternativa que ofreciese los mayores de todos. Al contrario: un decisor “pesimista” se centraría en los peores resultados, dando por supuesto que, tras cualquier decisión, siempre se presentaría la peor de las circunstancias posibles.

En cualquier caso, ¿existe alguien que sea 100% optimista o pesimista? De ningún modo, ni entre los gerentes de empresa ni entre el resto de la población. De ahí que el gran economista Hurwicz, aplicando este punto de vista, que nos parece trivial, propusiese calcular un “porcentaje de optimismo” y buscar soluciones intermedias, en lo que se conoce como criterio de Hurwicz. El gerente ya no sería “optimista” o “pesimista”, sino que estaría, como es natural, en el intervalo entre ambas situaciones. Otro ejemplo de “todo o nada” que se rompe.

“Fundamental Theory of Institutions: A Lecture in Honor of Leonid Hurwicz” de Roger B. Myerson

Talking about images I like… you may have read here mentions to “changing the landscape” to favour a desired result by making it natural for CoP participants to behave in the way you want.

In practice, you can’t forbid things that people want to do (and still keep them happy and active), and if you do forbid it, you can’t enforce it (CoP facilitators would need to be Sysiphus fans to enjoy it). If people’s incentives head some way, they will go that way. So maybe the solution is not to plant notices trying to divert them, but to change the incentives: offer different (and non-problematic) ways in which they can get what they want. Defuse the problem. Don’t change the rules, change the landscape.

In practice, that means changing the incentives and the available options. Find out what the drives are, and provide ways of satisfying them that don’t conflict with the CoP’s needs (compatible incentives). Build options that generate “convergent results”: ones that are good for the CoP and that incentivate the user.

Say that you run a set of online fora with volunteer facilitators (which means they want to dedicate as little time to chores as possible). Users will want to sell and buy things among themselves. But you want to avoid risky situations (fake sales and other creative ways of shortchange – which do crop up). Also, you don’t want merchants to set up shop in the forums for free (that’s what advertising is for). You can’t police every transaction nor mediate nor escrow (no resources). Simply forbidding trades hardly works, takes eternal explaining, and makes little sense for users.

So you change the rules. You provide a sheltered environment for trade in which only reasonably active users can enter, and erase every sale outside it (easy to set up, easy to maintain); this provides a built-in trust and reference screen, kills off web pirates who register just to trade, and gives users both a place to conduct sales among trusted people and an incentive to become active in the rest of the forums… plus, it makes it easy to screen grey merchants. Not everyone is happy – but active users and facilitators are. And the admin too, with no more frauds to worry about.

That’s a very down-to-earth example, but there’s loftier ones. This year’s Nobel Prize winner in Economics went to people (Leonid Hurwicz, Eric Maskin and Roger Myerson) who “laid the ground work” in defining “mechanism design” and played a very serious role in establishing it as a core part of Economics and a serious tool for political economists (and P2P application designers, and…).

“Mechanism” is a name for the rules and institutions that govern our economic activities – and you already know that I view CoP participation as a productive (and hence economic) activity. They realized that the design of the game (and specifically the incentives at play) affects the outcome, and have been delving into what improvements can be made to your plain-vanilla marketplace to improve some of its qualities (efficiency, fairness, you name it). Harnessing independent actors who act toward their own good… and getting a (better) common good as a result.

OK, more specific. Hurwitz defined a mechanism as “a communication system in which participants send messages to each other and/or to a ‘message center‘, and where a pre-specified rule assigns an outcome (such as an allocation of goods and services) for every collection of received messages”. To evaluate those outcomes, you have a goal function or set of criteria, external to the participants, although one of those criteria is usually to maximize the total utility won by participants.

Not all mechanisms are created equal, and the fun of the process is finding a mechanism (a set of rules, a communication mechanism) that always produces an optimal (goal-satisfying) result, whatever the options made by the participants.

For rules, read institutions (markets, other) and the incentives they hold to participants. If “participants” hints at a game, it’s because this is a branch of game theory… which are independent agents seeking to maximize their own utility. The rest of the concepts don’t need much explanation and take a lot of shapes.

That’s essentially it, without the maths. But maths help model all those possible mechanisms and quantify the probabilities of different types of results, thereby making these concepts usable beyond seat-of-the-pants implementations.

Some say that decisions and behaviours don’t depend on a rational evaluation of incentives and costs (and I quarrel with them), some say that CoPs are not markets (and I only say that they are a very convoluted transaction). But the relationship between the broad lines of the work of this year’s prize-winners, and day-to-day community design, seems quite clear . They’ve put the maths (lots of it) and the scope to an evident (but underused) idea, and are helping to improve the world.

Hats off to them. If you want to delve deeper, try this.

From the WSJ:

Americans Leonid Hurwicz, Eric S. Maskin and Roger B. Myerson won the Nobel prize in economics on Monday for developing a theory that helps explain how incentives and private information affect the functioning of markets.

For those who want it, here is more information about the men and their work:

• Scientific Background
• Leonid Hurwicz, Wikipedia
• Roger Myerson, homepage
• Eric Maskin, Wikipedia
• The Influence of Hayek on Hurwicz (HT: The Austrian Economists)
• Thoughts from Arnold Kling, Alex Tabarrok, and Tyler Cowen.
• Readings on mechanism design.

UPDATE: Differing perspectives from our friends in the Austrian school:

• Pete Boettke on the influence of Hayek on Hurwicz.
• Thomas DiLorenzo on subjectivism and market design theory.
]]>