Home > Uncategorized > Say ‘consistent’ one more time and I …

Say ‘consistent’ one more time and I …

from Lars Syll

Image result for say 'consistent' one more time memeBeing able to model a credible world, a world that somehow could be considered ‘similar’ to the real world is not the same as investigating the real world. The minimalist demand on models in terms of ‘credibility’ and ‘consistency’ has to give away to stronger epistemic demands. Claims in a ‘consistent’ model do not per se give a warrant for exporting the claims to real-world target systems.

Questions of external validity are important more specifically also when it comes to microfounded macro models. It can never be enough that these models somehow are regarded as internally consistent. One always also has to pose questions of consistency with the data. Internal consistency without external validity is worth nothing.

Yours truly has for many years been urging economists to pay attention to the ontological foundations of their assumptions and models. Sad to say, economists have not paid much attention — and so modern economics has become increasingly irrelevant to the understanding of the real world.

As long as mainstream economists do not come up with any export-licenses for their theories and models to the real world in which we live, they really should not be surprised if people say that this is not science.

To have ‘consistent’ models and ‘valid’ evidence is not enough. What economics needs are real-world relevant models and sound evidence. Aiming only for ‘consistency’ and ‘validity’ is setting the economics aspirations level too low for developing a realist and relevant science.

Say you are a die hard ‘New Keynesian’ macroeconomist that wants to show that the preferred ‘rigidity’ view on economy is the right one. Then, of course, it is a pretty trivial modelling matter to come up with what ever assumptions that makes it possible for him to construct yet another revised and amended DSGE model that is ‘consistent’ with his preferred view on the economy. But what’s the point in doing that when we all know that the assumptions used are ridiculously unrealistic? Using known-to-be false modelling assumptions you can ‘prove’ anything!

Economics is not mathematics or logic. It’s about society. The real world. And if you want to analyse and explain things in that world you have to build on assumptions that are not known-to-be ridiculously false.

Axioms of ‘internal consistency’ of choice, such as the weak and the strong axioms of revealed preference … are often used in decision theory, micro-economics, game theory, social choice theory, and in related disciplines …

Image result for amartya senCan a set of choices really be seen as consistent or inconsistent on purely internal grounds, without bringing in something external to choice, such as the underlying objectives or values that are pursued or acknowledged by choice? …

The presumption of inconsistency may be easily disputed, depending on the context, if we know a bit more about what the person is trying to do. Suppose the person faces a choice at a dinner table between having the last remaining apple in the fruit basket (y) and having nothing instead (x), forgoing the nice-looking apple. She decides to behave decently and picks nothing (x), rather than the one apple (y). If, instead, the basket had contained two apples, and she had encountered the choice between having nothing (x), having one nice apple (y) and having another nice one (z), she could reasonably enough choose one (y), without violating any rule of good behavior. The presence of another apple (z) makes one of the two apples decently choosable, but this combination of choices would violate the standard consistency conditions, including Property a, even though there is nothing particularly “inconsistent” in this pair of choices (given her values and scruples) … We cannot determine whether the person is failing in any way without knowing what he is trying to do, that is, without knowing something external to the choice itself.

  1. February 20, 2021 at 7:23 am

    Prof. Syll raised a very good question, which is also the focus of Algorithmic Economics. Let me explain how I solve it. Generally, a “consistent” model means that reasoning has been made to the “greatest” extent on which nobody “currently” can dispute. This implies hypothetically that reasoning can be made easily with null or negligible cost or time; However, when we insist on the significance of thinking time and cost consumed, what will happen? The answer is, we will “reasonably” allow the existence of conflicts and plurality in the system, because actors in the model would advocate that they have no enough time to solve the conflicts remaining inside, and they further commit that they will try to solve them in the future; besides, as decisions must be made timely in accordance with our introduction of time, the actors will have to bat out or fudge on thinking and calculation, thereby making imperfections, errors, and imaginations that arouse additional conflicts. Therefore, we can in a “consistent logic” to prove that inconsistences are the par for the course, except that the “inconsistencies” may become “consistent” in the future. The concomitance of conflicts with consistency in the space (where thoughts are treated as entities) and their possibility to cohere are called Algorithmically the “High-Order Consistency”. With this consistency, the model is still “consistent” despite inconsistent in the first order. On the contrary, that mainstream consistent model is really “unrealistic”, unreasonable (when introducing real thinking time) and then “inconsistent”! And, the “inconsistent” and “high-orderly consistent” model is nothing but the real world! We need not to, uneconomically, develop many other “inconsistent” models out of the real world, we just need to accept in principle the real world as a “model” and a prototyped “theory” on which we research and advise marginally. Many talks can be carried out afterwards. Thanks!

  2. Gerald Holtham
    February 22, 2021 at 6:08 pm

    An economic historian can weigh the facts of a given situation and, with a general view of human nature and what various actors were trying to achieve, together with a knowledge of the institutions and power-relations ruling in the time and place, she can give a convincing explanation of what happened. She can say why that happened and not something else. A great deal of wisdom can be accumulated in that way. A knowledge of relevant historical episodes can often guide us through current events, especially if we look for differences between then and now and think about how those differences will affect outcomes. Someone said (I forget who) history does not repeat itself but it rhymes.
    Economics tries to take a step further and identify the “rhymes”, that is elements that repeat themselves in numerous historical situations because they reflect persistent tendencies in economic life. When you try and extract such a tendency from real historical episodes you inevitably abstract from other things that were going on. Your abstract model will never fully explain or fit a real situation. Just like the person trying to draw a lesson from history you will have to look for the differences and adapt your ”model” to the situation being analysed. Economic theorists are looking for the general in the particular and tend not to worry about specific applications.(When they try to they often get it wrong). In any case the art of the practical economist lies in judging which model might be useful for a particular task and how to batter, complicate or distort it to make it work in context.
    It is possible to theorise well or badly, to produce models that might be useful in analyzing real situations or ones that will only mislead. It is easy to mock strong assumptions or abstractions made in constructing a model but it is surprisingly difficult to know in advance which abstraction will cut through complications and enable you to see a particular effect clearly and which one will take you somewhere completely pointless.
    One of the first abstract models in economic was Ricardo’s model of international trade. It could hardly have been more abstract, two countries trading two homogenous goods in conditions of full employment and constant returns to scale. From it he deduced the principle of comparative advantage which implies that countries can trade to mutual advantage even if one is more efficient that the other in producing everything. The conclusion was not obvious and it influenced British trade policy for the rest of the 19th century. This is typical of a useful economic model: there is no real world situation to which it corresponds; it is not sufficient to explain any trade relationship between two countries; there is a great deal of trade that is driven by other determinants and has nothing to do with comparative advantage. Nevertheless it is one of the influences going on in the real world, sometimes important, sometimes not. Isolating and explicating it was a scientific advance.
    Good economic models are like that. They are not photographs, more like a cartoon trying to capture a few striking features of a person. A good cartoon looks nothing like a real person, yet it makes them recognisable. To look at a caricature of General de Gaulle and say “but his nose wasn’t really that big” is to miss the point. I don’t “believe” any economic model but I have used a number in different contexts. Most contemporary macroeconomic models are a waste of time. But there is no practical alternative to modelling. As Becket said we just have to fail better.

  3. February 22, 2021 at 7:14 pm

    Gerry, when you write “Most contemporary macroeconomic models are a waste of time. But there is no practical alternative to modelling,” it makes me come to think of what Richard Lipsey (whose famous textbook was the first economics textbook I had to read when I started on my economics studies back in 1977) wrote in Journal of Economic Methodology (2001):

    “To get an article published in most of today’s top rank economic journals, you
    must provide a mathematical model, even if it adds nothing to your verbal analysis. I
    have been at seminars where the presenter was asked after a few minutes, “Where is your
    model?” When he answered “I have not got one as I do not need one, or cannot yet
    develop one, to consider my problem” the response was to turn off and figuratively, if not
    literally, to walk out. This reminds me of reports from seminars in Cambridge in the early
    1940s. The presenter was asked if he was assuming that savings always equalled
    investment and further attention to his presentation depended on whether he answered
    “yes” or “no”. What defines the “in group” of economists changes in each generation, but
    the requirement that to be listened to one must do what that group requires does not

  4. gerald holtham
    February 23, 2021 at 1:04 pm

    Lars, what you say is true. The economics trade is swept by fashion and does put an excessive emphasis on showing one has mastered the currently fashionable technique.. Indeed I have pointed out on this blog the paranoid biases in economics and speculated about the sociology of the trade that causes them. And you were good enough to agree with me. My plea is for more pluralism and less factionalism. So let me qualify the remark “there is no alternative to modelling”. I did not mean modelling is or should be the only form of useful work. Of course useful work can be done without a formal model. Insights are worth communicating before they can be formalised. The more you develop the insight, however, exploring the conditions under which it holds and what its limitations might be, the more you approach a complete specification that can be tested. I do believe that follow-on process and the empirical testing is necessary for knowledge to progress. Perhaps this is too instrumentalist a view for some but surely a discipline reaches a peak when it is specific enough to be useful in designing policies that promote human welfare.
    We both deplore the turn that macroeconomics has taken but seem to have different diagnoses of the root of the problem. You blame excessive formalism. There is some of that but I put more responsibility on the wrong sort of formalism coupled with a disdain for empirical work that rejects what is in fashion. You see econometrics as part of the problem but the mindset of the econometrician is radically different from that of the ideological theorist and I see it as part (though only part) of the solution.,

  5. February 23, 2021 at 3:25 pm

    As a mathematician, I find this kind of thing very confusing. In mathematics the consistency of a model is a separate consideration from its interpretation. Consider:

    “Can a set of choices really be seen as consistent or inconsistent on purely internal grounds, without bringing in something external to choice, such as the underlying objectives or values that are pursued or acknowledged by choice? …”

    One way to ‘save’ Samuelson’s axioms for the example given is to suppose either that having a reputation is a ‘good’ or that losing one’s reputation is a ‘price’. So when economists talk about the consistency of ‘the model’ they seem to mean consistency between the model and its interpretation. If this is the concern, why not be more explicitly concerned about ‘model theory’, rather than abandon mathematics on the grounds that economists often use ‘mathematical models’ that are inconsistent with their interpretations?

  6. February 24, 2021 at 6:17 pm

    I do not understand why Lars Syll has cited the quotation in the last part of his article with no mention on the author and the source. From the picture, I came to know that it was cited from Armartya Sen’s Rationality and Freedom (Harvard University Press, 2002: Paperback 2004).

    What kind of logical relation does there exist between the cited part and the argument Lars Syll presented in the main part of this blog article? Sen studies (or better questions) the internal consistency of choice. The first cited paragraph explains Sen’s motive to question the internal consistency of choice. The whole paragraph reads:

    Internal consistency of choice has been a central concept in demand theory, social choice theory, behavioral economics, and related fields. It is argued here that this idea is essentially confused, and there is no way of determining whether a choice function is consistent or not without referring to something external to choice behavior (such as objective, values, or norms). We have to re-examine the robustness of the standard results in this light. The main formal results presented here is an extension of Arrow’s General Possibility Theorem. This drops the need to impose any condition of internal consistency of social choice, or any internal notion of “social rationality.”(pp.122-3)

    Obviously, Sen is questioning the traditional treatment of theory of choice and claiming that it is necessary to refer to “something external to choice behavior”. His conclusion in Lars Syll’s quotation “We cannot determine whether the person is failing in any way without knowing what he is trying to do, that is, without knowing something external to the choice itself” corresponds to the closing sentence of my quotation.

    Sen does not argue the logical inconsistency of a macroeconomic model nor the external validity of the model. Simply stated, Sen is arguing a totally different problem than that Lars Syll argues. Citation from Sen does not confirm anything that Syll contented. I am afraid that Lars Syll is extremely confused.

    • February 24, 2021 at 7:39 pm

      Author and source are given in the original post on my blog (and I don’t know why that has fallen away here). Mystery solved :)

      • February 25, 2021 at 7:18 am

        Thank you for the reply. It would be more convenient for other readers that I give the link to the blog article:

        The question of “author and source” is solved. But, the mystery of logical relations between the citation from Armartya Sen and the main part of the article is not solved yet. There is no explanation in the article nor in your above comment.

    • February 25, 2021 at 2:21 pm

      My blog post argues that CONSISTENCY requirements are not enough to give models export warrants. Sen discusses CONSISTENCY axioms in rational choice theory . And you fail to see the connection? You have to be joking, Yoshinori!

  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.