Post hoc, ergo propter hoc …
Although causality seems such a simple concept, it is notoriously difficult to prove conclusively, particularly in a social science discipline such as economics. James Tobin, a Yale economist and Nobel Laureate, published a paper in 1970 with the same title as this post, in which he demonstrated that you cannot infer causal relationships by observing the ordering of turning points in economic variables alone. To claim that temporal priority necessarily means causal priority is to fall victim to the post hoc fallacy.
Literally translated the phrase means “After this, therefore because of this” although I prefer the more succinct version -- “After, therefore because of it” -- given by Leo McGarry, the White House Chief of Staff in the second ever episode of “The West Wing”. There is some banter in the Oval Office between President Bartlett and the Press Secretary C.J. Cregg about their failure to capture Texas in the election being caused by the President's sense of humour. My edited version of the conversation is as follows.
C.J. Cregg: When asked about campaigning in Texas, you say you look bad in funny hats.… Point is, we got whomped in Texas.
President Bartlett: C.J. on your tombstone it's going to say “Post hoc, ergo propter hoc".
C.J. Cregg: Okay, but none my visitors will be able to understand my tombstone ...
(McGarry gives the definition)
President Bartlett: We did not lose Texas because of the hat joke. You know when we lost Texas?
C.J.Cregg: When you learned to speak Latin?
President Bartlett: Go figure.
The issue of causality is somewhat different at the level of individual economic agents. Here natural experiments, that allow for the seemingly random assignment of study subjects to different groups, can help to identify the cause and effect of policy changes. Even in these situations, however, there are severe technical difficulties that can hinder accurate inference.
So what has all this to do with life at Cracroft Farm? The glorious autumn weather this week has been conducive to being outside, and while working away I got to thinking about causality in econometrics in a desultory sort of way (a bad habit of mine). Having spent a lot of time in my academic career pondering on this vexed question, you can (perhaps) forgive my musings on this subject.
I observed that there seemed to be many wonderful examples of causality in action in the garden! Just take a look at the domestic orchard for example. Standing at the south-west corner, I look left and I see unkept paddock; taking a half-turn right, I see an oasis of green. The causality here is quite clear. If you build a fence to keep out the local fauna, provide water and cut the grass regularly, you will end up with something half decent. There are no confounding effects of microclimate to offer possible alternative explanations. The fence came first and then the orchard thrived.
Feeling as though I had finally cracked the post hoc conundrum, I strolled around looking for more manifestations of causality in nature. The first thing I came across was one of our wonderful birch trees. We bought this tree (along with two others) in December 2020 for $2, when it was not more than 10 cm high. Its growth has been nothing short of miraculous and its autumn show this year is fantastic. Here was another example: autumn comes, and the leaves turn.
Alas, my new found certainty was then unceremoniously shattered by our ornamental pear trees. Autumn is nearly done and there is hardly a coloured leaf to be seen on any of them. Moreover, driving around the Huon I now notice that everywhere there are trees of the same variety that are showing completely different degrees of autumnal progression. Clearly autumn’s arrival does not necessarily mean the same thing to all deciduous trees, even when they are of the same species and planted right next to each other. There are obviously still some confounding elements that inhibit our understanding of causality and autumn foliage.
There is, however, one situation where I can guarantee that the direction of causation will be unequivocal. This weekend a major cold weather system – dubbed a “polar surge” – is forecast to sweep across the south and southeast of Australia, including Tasmania, bringing heavy rain, damaging winds and snow. When this polar surge finally arrives, Pepper and I will definitely retreat indoors where I will sit in front of the fire while Pepper takes the sheepskin. There is absolutely no need to worry about the post hoc fallacy in this instance ...
Comments
Post a Comment