…the development of science is driven, in normal periods of science, by adherence to what Kuhn called a ‘paradigm’. The functions of a paradigm are to supply puzzles for scientists to solve and to provide the tools for their solution. A crisis in science arises when confidence is lost in the ability of the paradigm to solve particularly worrying puzzles called ‘anomalies’.
https://plato.stanford.edu/entries/thomas-kuhn/
(Trivia question for econ nerds: what long running series in an academic journal are you reminded of when you read this? I’ll answer the question in tomorrow’s post.)
I was one half of a very nice conversation we recorded over the weekend for a podcast. The podcast will be out in a couple of weeks or so, and I will write more about it then. But for today, we’re going to talk about a lovely little essay that I read after many, many years. I re-read this essay because it came up in conversation during the podcast.
The title of the essay is “Economics as an Inductive Science“, and it is the Presidential Address that was delivered at the sixty-third annual meeting of the Southern Economic Association, in 1993. It was delivered by Robert Clower. If you’ve read enough macroeconomics in your life, Clower is a name that is bound to come up sooner or later. But this particular essay isn’t necessarily (or exclusively) about Keynesian economics. It is, instead, a rant about where economics seems to be headed. And it is a rant that is worth reading.
Apart from some kickass quotations (“Logic is the art of going wrong with confidence” is the first quote, and it only gets better from there on in), it is an essay that makes you think about how we go about “doing” economics.
Economics (and especially macroeconomics), Clower says, is riddled with Kuhnian anomalies. What are Kuhnian anomalies? That’s why we began with that quote at the top of the blogpost.
But let’s back up for minute. What do we do as economists? We learn principles, theories and models (and some of us even develop some of these principles, theories and models). We then go look at the world. Based on what we see in the world, and based on what we think we know of how the world works, we make certain predictions.
Note that the word prediction doesn’t necessarily mean forecasting the future. “If we see x going up, y should be going down” is a prediction, but not a forecast.
Now, if these predictions turn out to be repeatedly and wildly wrong, we are left with three hypotheses. Either, we should reason, the data we have collected is wrong. Or our theory is wrong. Or both! And then, as “scientists”, we go and take a look at our data collection methodologies. Once we ascertain that we haven’t erred there, we look at the theory. And if the data looks right but the theory ain’t working out… well, then, we have an anomaly.
Collect enough of these anomalies, and you get to challenge theory. And every now and then, we get to update the theory. And so science chugs along, inch after excruciating inch.
If you are looking for a concrete example of how this works in practice, I have a story to tell you. This story consists of a lot of things, including static on your television set, a Nobel prize in physics… and pigeon poop. Yes, really.
“I had a lot of experience fixing practical problems in radio telescopes,” Robert Wilson now says. He and his wife Betsy Wilson still live in Holmdel, New Jersey, not far from hilltop where the tests were run. “We looked for anything in the instrument or in the environment that might be causing the excess antenna noise. Among things, we searched for radiation from the walls of the antenna, especially the throat, which is the small end of the horn. We constructed a whole new throat section and then tested the instrument with it.”
https://www.smithsonianmag.com/smithsonian-institution/how-scientists-confirmed-big-bang-theory-owe-it-all-to-a-pigeon-trap-180949741/
At one point, new suspects emerged. Two pigeons had set up housekeeping inside the guts of the antenna. Maybe their droppings were causing the noise? Wilson and Penzias had the birds trapped and then cleaned the equipment, but the signals continued.
After a year of experiments, the scientists concluded that they’d detected the cosmic background radiation, an echo of the universe at a very early moment after its birth.
“We started out seeking a halo around the Milky Way and we found something else,” notes Dr. Wilson. “When an experiment goes wrong, it’s usually the best thing. The thing we did see was much more important than what we were looking for. This was really the start of modern cosmology.” In fact, Wilson and Penzias were awarded the Nobel Prize in Physics in 1978 for determining that the hiss they were hearing wasn’t pigeon poop at all, but the faint whisper of the Big Bang, or the after glow that astronomers call the cosmic microwave background.
You set up an experiment, and you expect to see x. “Expect” to see x? That’s the current theory!
But hey, you don’t see x, you see y instead. Well, maybe we’re seeing y because there’s some problem with our data gathering apparatus. That is, we’re sticking to our guns when it comes to our theory, and we’re going to see if our data collection strategy is off. But if you eliminate every single source of data contamination (including pigeon poop!), then it is finally time to challenge established theory and update it – and win a Nobel Prize to boot.
Cool, huh?
Well it is, except for the fact that Clower is really telling us that this doesn’t happen enough in economics. Consider this quote from Einstein, cited in the speech we are talking about today:
“Science is the attempt to make the chaotic diversity of our sense-experience correspond to a logically uniform system of thought. [ … ] The sense experiences are the given subject matter, but the theory that shall interpret them is man-made. It is the result of an extremely laborious process of adaptation: hypothetical, never completely final, always subject to question and doubt”
Clower is making the point that our theories have become completely final, and are not subject to question and doubt. Read in particular Clower’s descriptions of his set of stylized facts (pp5), and the facts as assumed by our paradigmatic model (pp6).
As economists, we have two choices:
We can keep searching for the pigeon poop, even though it may not exist. Or we can ask if our theory needs to be updated. And maybe, Clower suggests, we should work on updating our theories a bit.
And while doing so, he suggests keeping three things in mind.
While doing inductive science, he says, one’s motto ought to be that “If it isn’t common sense, it’s probably wrong.”
Second, a direct quotation from the speech: “If we are ever to be taken seriously as scientists we would be well advised to proceed with this task as most practitioners of other inductive sciences have proceeded – by taking a hard look at the world around us in a serious effort to lend intellectual order to the “chaos” that strikes our eyes at first sight”
And finally, an admonition of sorts. We should remember, he says, that inductive sciences deal with plausible inference, not with demonstrative reasoning.
EconForEverybody 