Discourse in Wonderland
Imagine three groups of astronomers who have differing models of how fast a distant galaxy is expected to grow.
Group 1 predicts growth of 20% over the coming year.
Group 2 predicts growth of 30% over the coming year
Group 3 predicts the galaxy will shrink by 10% over the year.
So a year passes, the galaxy shrinks by 10% and a leading newspaper runs a story about astronomers predicting galaxy sizes. Does the story tell us that:
A) Groups 1 and 2 are meeting with group 3 to better understand how group 3 predicted what they missed so they can adjust their models accordingly.
OR
B) Group 3 has made their prediction of next year's growth of the galaxy (10% more shrinkage) which is considered newsworthy given the accuracy of their last prediction.
OR
C) Groups 1 and 2 are meeting with each other over a dispute between them because Group 1 is predicting the galaxy will grow 25% for the coming year, while Group 2 is predicting the galaxy will grow by 40% over the coming year. Neither has changed their model since last year and Group 3 is not mentioned.
If you picked C) then you understand how our public discourse works.
Take a story in today's Globe and Mail, for example:
A few comments:
1) First of all, everybody knows that all of these economic forecasts are worthless, so the entire thing is just a big hoax. Not a deliberate hoax as much as everyone just playing along in a game they know is meaningless. Disagree? Then why doesn't the story even mention the accuracy of past forecasts by the two groups? Because they know that the accuracy of the forecasts is irrelevant, that's why.
One of the interesting things about moving from doing math in elementary school and high school to doing it at university was the transition from a world where every year you learn more about all the things we know about math (elementary/high school) to a world (university) where every year you learn more about what we don’t know, what we can’t know and how we know that we can’t know it. Complexity theory, chaos theory, Incompleteness theorems – these concepts have the names they do for a reason.
Sadly, predicting the total value of transactions involving money that we will make in 2010 (i.e. GDP) is one of those things that has proven itself to be not particularly amenable to being modelled mathematically. So far, applying the lessons of history (big debt bubbles cause trouble when they burst) and thinking things through has been much more fruitful as a method of analysis and prediction (for example, see here).
2) Is it just me, or is there something crazy in that the Bank of Canada is meeting with economists from 11 other banks to talk about this. Are there no economists in Canada who don't work for a bank? What about the provincial governments, large (non-financial) corporations, the universities – all of these must have economists on staff, right? I find the dominance of 'bank' economists in Canada very puzzling.
3) As Stephen Gordon points out, there’s not much difference between the two groups anyway.
4) As the first comment on Stephen’s post points out, given that so far both groups have consistently erred on the high side since the downturn hit, continuing to believe they are too optimistic seems reasonable at this point.
5) This is all symptomatic of our accountability free media world where getting heard depends on whose interests you serve, and being right about things is irrelevant. And no matter how wrong you are, or how often, there's zero consequences.
Group 1 predicts growth of 20% over the coming year.
Group 2 predicts growth of 30% over the coming year
Group 3 predicts the galaxy will shrink by 10% over the year.
So a year passes, the galaxy shrinks by 10% and a leading newspaper runs a story about astronomers predicting galaxy sizes. Does the story tell us that:
A) Groups 1 and 2 are meeting with group 3 to better understand how group 3 predicted what they missed so they can adjust their models accordingly.
OR
B) Group 3 has made their prediction of next year's growth of the galaxy (10% more shrinkage) which is considered newsworthy given the accuracy of their last prediction.
OR
C) Groups 1 and 2 are meeting with each other over a dispute between them because Group 1 is predicting the galaxy will grow 25% for the coming year, while Group 2 is predicting the galaxy will grow by 40% over the coming year. Neither has changed their model since last year and Group 3 is not mentioned.
If you picked C) then you understand how our public discourse works.
Take a story in today's Globe and Mail, for example:
"At a meeting likely to last five hours at the central bank's headquarters in Ottawa, the economists are expected to spend the first 30 minutes duking it out with central bank officials over their divergent views of Canada's prospects.
Following opening remarks by deputy governor John Murray, the first item on the agenda is a discussion on "forecast comparisons and issues arising" from the Bank of Canada's latest policy update and "chief economists' forecasts," according to a copy of the agenda obtained by The Globe and Mail.
Present will be senior policy makers and chief economists from at least 11 of Canada's biggest lenders.
The issue is Mr. Carney's prediction last month that the recession will begin to recede by the end of the year, followed by a sharp rebound of growth averaging 3.8 per cent in 2010. While Bay Street accepts the premise of a recovery next year, the bank analysts expect growth at only about half that pace."
A few comments:
1) First of all, everybody knows that all of these economic forecasts are worthless, so the entire thing is just a big hoax. Not a deliberate hoax as much as everyone just playing along in a game they know is meaningless. Disagree? Then why doesn't the story even mention the accuracy of past forecasts by the two groups? Because they know that the accuracy of the forecasts is irrelevant, that's why.
One of the interesting things about moving from doing math in elementary school and high school to doing it at university was the transition from a world where every year you learn more about all the things we know about math (elementary/high school) to a world (university) where every year you learn more about what we don’t know, what we can’t know and how we know that we can’t know it. Complexity theory, chaos theory, Incompleteness theorems – these concepts have the names they do for a reason.
Sadly, predicting the total value of transactions involving money that we will make in 2010 (i.e. GDP) is one of those things that has proven itself to be not particularly amenable to being modelled mathematically. So far, applying the lessons of history (big debt bubbles cause trouble when they burst) and thinking things through has been much more fruitful as a method of analysis and prediction (for example, see here).
2) Is it just me, or is there something crazy in that the Bank of Canada is meeting with economists from 11 other banks to talk about this. Are there no economists in Canada who don't work for a bank? What about the provincial governments, large (non-financial) corporations, the universities – all of these must have economists on staff, right? I find the dominance of 'bank' economists in Canada very puzzling.
3) As Stephen Gordon points out, there’s not much difference between the two groups anyway.
4) As the first comment on Stephen’s post points out, given that so far both groups have consistently erred on the high side since the downturn hit, continuing to believe they are too optimistic seems reasonable at this point.
5) This is all symptomatic of our accountability free media world where getting heard depends on whose interests you serve, and being right about things is irrelevant. And no matter how wrong you are, or how often, there's zero consequences.
Labels: economists, lack of accountability is killing us, replace the math with stories
posted by Declan at
Wednesday, February 04, 2009
3 Comments:
For what it's worth, there's been some rough work done on the relative projections of different types of economists in the U.S. - with banks somewhere in the middle, albeit with generally more optimistic projections than non-profits and academic economists.
By
Greg Fingas, at 6:48 AM
Thanks for the link. More of the same nonsense, comparing forecasts to forecasts rather than to reality - but interesting nonetheless
How hard would it have been for them to have a second chart showing the same predictions from a year ago - alongside what actually happened?
By
Declan, at 4:49 PM
Declan,
you need to remember the economic forecasters defense:
Hey meterology is a science and it is never very accurate fourteen days ahead of time and they constantly revise their predictions!
Of course meteorologists forecasts are of a very different kind. First they do a pretty good job at predicting the weather within fourteen days. Second, they get the daily right and don't claim much for the medium to long term. Third where they do get further out they tend to say things like a wetter summer or colder winter. They do not say things like "we are predicting 2.3% more preciptation for next summer" or "a 1.6 percent colder winter."
Economic forecasts work-off the publics' willingness (or perhaps just economists) psychological desire for scientific certainty.
It helps to simply treat economists forecasts more like an opinion pole: the opinions expressed need not bear any relation to reality.
By
Anonymous, at 9:59 PM
Post a Comment
<< Home