You are viewing a read-only archive of the Blogs.Harvard network. Learn more.

The Longest Now


Field failure: macroeconomics
Saturday October 04th 2008, 11:32 pm
Filed under: Uncategorized

Macroeconomics is less than a century old.  It has suffered from a lack of predictive power and difficulty reconciling its ideas with the more testable and tangible ones of its micro cousin.  Its claims and schools have guided many global scale political, military, and financial changes.  Like man, I imagined since college that it was simply passing through field adolescence; only recently have I seen it as suffering from repeated field failure.

What does it mean for an entire field to fail?  You can believe that the universal elements are fire, air, water and earth, and generate useful elaborate rules of thumb with some predictive power.  You can found a field on the notion of spontaneous generation, including crocodiles being generated from rotting logs, and both cultivate increasingly-refined correlations and produce testable lemmas and hypotheses starting with these ideas.  Some of these hypotheses can be proven — a hypothesis inspired by false assumptions can well be true — further developing the field.  But when extremes or boundary events cannot be explained by a field’s axioms; when the founding principles of a field cannot withstand scrutiny; or when there is no macroscopic description of the world consistent with its concepts : then it slowly becomes clear the field has failed.

No field is ever acknowledged as failed within a generation; even when a breakthrough or mistake is so tremendous as to shake it to its foundations.  But individuals do recognize it, and repeated observable failures lead to rapid change and growth of understanding.  Macroeconomics has experienced many minor setbacks and confusions in the past, sometimes written off as a result of the field’s youth.    The differences between different forms of capitalism, orcapitalism v. socialism, or what have you are coming to seem like approaches that are appropriate for different circumstances, not foundational differences between two religions, one of which is Always Right and the other which is Forever Wrong.  

What have been viewed to date as philosophical differences in approaches to ‘economics’ and ‘growth’ — and I note the notion of philosophy or dogma as a valid determinant of theory (rather than nuanced experiment) may be a characteristic of a field not yet capable of success — are looking less and less that way.

I put ‘economics’ and ‘growth’ in quotes there because the modern concept of economics is generally stripped of much of the social and moral elements that made up its description one or two centuries ago; and as such feels like a regression to me.  Certainly our modern notion of growth is hollow, short-sighted and immature compared to concepts of growth in other places and times; but has been paired with powerful forces of social, communicative, and innovative development so that few would even think to separate one from the other.

Each day I encounter one or two serious conversations which are crippled by mundane conceptual failures in economics as a broad field. Conversations about how to direct effort, assign importance, focus vision or evoke change; about how to govern organizations and distribute tasks for great Projects.  These conversations are hampered by a scarcity of variation in ideas about give and take, about bounty and exchange; about resources, production, opportunity, supply, moderation, sharing, desire, demand, need, contingency.   They are hampered by the aggressive spread of a few simplistic rules of thumb, and the confident assurance that those are both true laws and the only ones.




I’m an engineer. I work on control systems. For most control problems we have a variable, like the temperature in a room, a set point, like the temperature we would like it to be in the room, and a controlled item, like a radiator or an air conditioner.

We vary the controlled item – the heat given off bay the radiator based on the current radiator setting and the difference between the current temperature and the desired temperature. We understand how to control the radiator so the room temperature gradually comes up to the set point. We understand how changing a parameter can mean the temperature overshoots the set point and goes way high then comes back down and overshoots again and then wobbles above and below the set point and never settles down.

For the economy we have two controlled variables – interest rates and taxes. We have two variables – inflation and unemployment. This is not a dificult control problem. The problem is our politicians have figured out that if they can make inflation and unemployment wobble widely then the highs and lows are more extreme so they try and ride the tiger and get inflation and unemployment both low at election time. Time to adjust their incentives.

Comment by joe 10.05.08 @ 5:11 am

“For the economy we have two controlled variables – interest rates and taxes. We have two variables – inflation and unemployment.”

I don’t think it is that simple. Economies are first built on foundational trust (in currency, in continuity of things such as “interest rates” and “valuation”), then on the availability of trusted and understood market interactions (including “credit”, “debt”, access to specific currencies [each with its own trusted fairly-continuous value] and other forms of liquidity). Then you can start to talk about internal interest rates, taxes, and the quantity of currency and government bonds in circulation.

As for relevant variables and measures, unemployment is very tangible and does not depend on these layers of trust; inflation and valuation of currency is more abstract and does depend on some of them. Other variables such as the going payday-loan interest rate are also important to certain groups, and not directly controlled by any central group.

When you get to the truly macro level, and are dealing with the entire world, you’ve lost if your method of defining and accounting for value misses any major class of production. Which any system that relies solely on currency exchanges must. What is the scale of this oversight? I would guess it is a few orders larger than what is currently measured. So our best guesses as to what maximizes or minimizes ‘growth’, ‘productivity’, or ‘externality’, using these measures, is rather unlikely to have any bearing on reality. (Just as a poll of a small randomized group can tell you about the leanings of a group many orders larger, one can hope that the subset of value captured by current measure-systems is sufficiently well-distributed across the different types of value that we have some idea of how well we are doing. But that may well be wishful thinking.)

Comment by metasj 10.06.08 @ 3:18 am





Bad Behavior has blocked 191 access attempts in the last 7 days.