Teaching Economics – You Need to Teach Contested Key Ideas Before You Can Model

How to teach Economics has become a major issue.  The Bank of England has sponsored an INET conference on how to teach economics post-crisis.

As readers of this blog will no i’m not an economist, its merely a hobby; but I suspect that even if I had unlimited resources to study or my life over again the last thing I would want to do is take an economics major.  From my own readings I know that modern economics teaching reinforces too many errors, has become isolated from mainstream thinking and theory in the social sciences, and worst of all often has a contempt for its own history so that the history of economic thought is often an optional module, not considered important, and worst taught in the worst whiggish form as a series of errors building to a triumphant conclusion.

What kind of test do we need to set for what is a good economic curriculum?  I would suggest it is producing the sort of graduate that other graduates would want to speak to and take advice from in answering open questions about history and current and anticipated problems.  Too often the answer is with rational actors and an unregulated market that problem cant happen, a none answer which means that the questions don’t get asked.  A sign of health for economic departments is when other departments are banging on their doors to conduct interdisciplinary research.

I don’t agree however with the track INET is going down of a mainly empirical  ‘teach macro first’ course.  The problem as Noah Smith points out is that Macro is the one part of economics that no-one can agree on.

It is important that Economics be founded on an understanding of current problems and trends, the problem with a largely empirical and/or historical course is it does not  necessarily provide a good foundation for modelling.

There seems to be, rightly, a reaction against courses being solely about teaching the skills at undergraduate level necessary to do DGSE type modelling at post-graduate level.  This will simply dissillusion students who don’t go on to postgraduate study and teach them nothing but ideology rather than how to critically apply economic ideas.

However equally it is a danger to revert to a solely marshallian partial equilibrium type course.  In a widely cited blog post Ryan Decker explores the dangers of partial equilbrium thinking and the advantages of proper modelling. 

In typical empirical work and heuristic/narrative theorizing, it’s really difficult to avoid partial equilibrium reasoning. A DSGE model, even a very simple one, has more moving parts than my mind alone can keep track of, and it forces agents to obey resource constraints in a way that is really difficult without formality.

I don’t agree with him that DGSE is the best form of modelling.  It has largely got is into the problem we are in.  However as Ramanan says in response.

Personally I find stock-flow consistent models extremely useful to think about the working of the world as a whole. It is when you sit down and work through the models, that you realize how complicated the whole thing is and how naive intuition isn’t good enough.

DGSE models in ignoring state (stocks) and treating all agents with a single representative agents, rather than differential receivers of factor incomes as distinct agents were unable to properly model financial intermediation, money, liquidity or debt.  But a properly education future generation of economists must be able to understand and model these in ways which are rigorous in their accounting.

So what is the role of economic history in building such understanding?  The history of economic thought is too big a subject to a taught in a one semester course in the first year.  All that can sink in from that is a chronology of economists.  Rather I think it should build over several years into an inquiry of open questions in economics.  Indeed I don’t think any of the great economists can be properly thought of in terms of providing partial and fully correct discoveries.

To my mind an economics curriculum needs to be founded around these contested core ideas, ideas such as what is interest? what are profits?  what is saving and what is the impact of saving?  What creates value?  What is capital?  There is no real agreement on any of these ideas.

To take just one example – savings.

Such a course might start off with the classical Turgot/Smith theory of savings and capital accumulation.

Then you would move on to the concept of the ‘paradox of thrift’ Mandeville,and Roberston.

Then the early Austrian reformulation of  the classical theory by Bohm-Bawerk where savings are intertemporal consumption.

Then the debate triggered by Bostedo’s criticisms of Bohm-Bawerk that he conflated two things (capital accumulation and not spending) and ignored the paradox of thrift, and Bohm-Bawerks famous response that money not spend adds to bank funding (implicit is loanable funds).

Then Keynes reformulation of the ‘paradox of thrift’ based on the savings=investment identity and Kaldors dynamic proof of this and the spending multiplier.

Then  briefly dealing with the Ramsey theory of the optional savings rate.

Then finally all of these ideas would be revisited in terms of what we know now about money creation and accounting.  This would reveal that none of the thinkers above provided full closure, that is because although investment creates savings ex-poste that investment if provided through loans still requires an expansion in banks capital ex-ante  financed through profits invested from other firms.    Also an endogenous money banking theory perspective (such as Davenport and Phillips) shows how excess reserves can propagate expanding banks potential lending even without creation of state money.

If studied alongside other core contested concepts such as interest the final year students would be able to understand the different ways in which these modules could be fitted together and how models could be built.  It would require not a great degree of mathematical understanding to appreciate these modules.  Rather the greater the mathematical understanding the greater options the scholar has in building the models.  Such models would not be apriori general equilibrium, but once a model was comprehensive and properly covered stock and flow and all factor returns it could be used to test whether GE is achieved and behavior during disequilibrium.  Much more useful.

Would not students taught in this way not be better modellers?  Yes because in understanding the core concepts and debates they could model and test competing theories against different assumptions.  The true mark of a science.


9 thoughts on “Teaching Economics – You Need to Teach Contested Key Ideas Before You Can Model

  1. Good points.

    You said, “but I suspect that even if I had unlimited resources to study or my life over again the last thing I would want to do is take an economics major. ” I couldn’t have said it better my self.

    There seems to a few sides. The teaching, the learning, and the student investment.

    Here is a formula for students and potential students.

    As the limit of economics education approaches zero = get a degree in some thing else.

    E-Con education or as an art, isn’t worthy of students time, investment, or youth.

    It’s worth is negative. Unworthy of investment, let alone a large investment.

  2. Just a couple of points of clarification: First, I did not argue that “DGSE is the best form of modelling.” In my post, I tried (and apparently failed) to suggest that a variety of approaches to economic analysis can be useful. The post was only meant to defend an approach that I have found useful.

    Also, if this is an important topic to you, I would encourage you to explore mainstream macro literature a little bit more. DSGE is not equivalent to “representative agent.” All models are simplifications along some dimensions, and it’s true that some DSGE models simplify along the dimension of human heterogeneity, but there is a wide variety here. A lot of people (including me) work with DSGE models that have agents that are heterogeneous along a variety of dimensions–wealth, time preference, labor productivity, entrepreneurial ability, etc. Additionally, DSGE models don’t ignore stocks/states. It’s true that they often don’t focus on stock/flow as much as some other approaches, but almost every DSGE model tracks the state of households, firms, the economy, etc. Finally, DSGE models do not have automatic political implications. They can be used to make cases in favor of regulation and other government interventions. It’s a very versatile tool. In fact, a lot of the economic concepts you describe in this post can be comfortably modeled in a DSGE setting.

    This is not meant to convince you of their usefulness–but I think some of your argument here is addressing an inaccurate caricature of the DSGE modeling approach.

  3. Well, Rosser is ok. He doesn’t really get to the purpose of het agent models. A lot of people have used these to show things that the rep agent models get wrong (eg Heathcote 2005), while Rosser is suggesting that they end up just behaving like rep agent models.

    Once again, I think you miss out on the range and variety of DSGE models. Some people use cash-in-advance, yes, but there are huge portions of the literature that introduce money in other ways, like search.

    I will look at your other links, but I must say that I (and maybe others) would take your criticisms more seriously if you demonstrated more familiarity with DSGE literature. That is not to say that familiarity is a precondition for criticizing, of course–only that some of your very specific criticisms do not accurately describe the literature. I understand the case against DSGEs. I understand that they are oversimplified, rely on many assumptions, and lack predictive power. I understand that they are not the only path to economic insight or measurement. What I don’t understand is the confidence of their critics, who seem certain that they have alternatives that are clearly better. My prior is that macroeconomics is too complicated to justify that kind of attitude (and yes, I know that some DSGE advocates share the attitude). I have yet to see a modeling approach that will significantly move that prior.

    • I have failt good familiarity with the DGSE lit however your request is rather like asking a surgeon who has adopted a new technique to be more aware of the literature on leachs. There is no need if better techniques give palpably superior results, and adding ‘search’ as a means of modelling money only makes my point on this. Ok a challenge to solve it – take this one of my recent models https://andrewlainton.wordpress.com/2013/07/26/a-simple-post-keynesian-alternative-to-is-lm/ and try a) to model the double entry boookeeping in DGSE 2) model it in SFC. The second is elementary, the first is a nightmare. Then you can decide for youself which is superior.

  4. Pingback: Teaching Economics? Start with Key Contested Ideas | Unlearning Economics

  5. Pingback: ‘Savings’ can Increase Real Wages: A Reply to @asymptosis | Decisions, Decisions, Decisions

  6. Pingback: Towards a Formally Defined New Economics | Decisions, Decisions, Decisions

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