Ryan Avent recently made a very astute comment:
In my new book (of which I just received a copy), I argue that when considering demand shocks to the economy we should think of causation in terms of policy counterfactuals. Thus if the lowest cost way of preventing the Great Recession would have been to set monetary policy in a different (more expansionary) position, then it makes sense to argue that tight money caused the Great Recession. As an analogy, if a bus driver could have avoided going off a cliff by setting the steering wheel at a different position, then it makes sense to argue that inept steering by the bus driver caused the accident (not a turn in the road).
READER COMMENTS
Phil H
Aug 23 2021 at 3:13am
I think I absolutely disagree with this in public policy. It might make sense in questions of legal liability, but for public policy this is the wrong way to think. I’m much more inclined to think about the stories of how aircraft safety was improved.
I can’t find the reference now, but I remember one early story was that there were a surprising number of fighter pilots landing without putting their landing gear down. The designers realised that the landing gear switch was very close to another similar switch; they changed the design; the problem went away. For each individual landing, it would be reasonable to say that the failure occurred because the pilot did not pay enough attention. But that doesn’t (didn’t, wouldn’t) solve the problem. The only way to prevent future failures is to change the design.
Without understanding the institutional background, I don’t know if the great recession was a case of one pilot making an error, or many pilots repeatedly making errors. But seeing as similar problems recurred following 2008, it seems like this is a systems problem. So while it may be reasonable to blame the “pilots” who messed up monetary policy in the 1930s, it may not be reasonable to think that this is the best way to fix the problem.
Jon
Aug 23 2021 at 7:31am
Phil H,
I think you’ve conflated two separate things from aircraft safety history, causing you to miss that Scott’s point is also directly applicable to your counter example.
Landing gear up was not caused by the switch location, but instead this switch location issue would cause issues by pilots accidentally retracting the gear in the post-landing roll out, when pilots were frequently reaching to retract the flaps (a standard post landing action) but accidentally moving the gear switch (obviously not a desirable post landing action).
In contrast, landing gear up was caused by pilots just plain forgetting to lower the gear, and not from switch location.
However, BOTH undesirable results (landing gear up and accidentally retracting the gear after landing) were reduced by increasing professionalization of the cockpit (e.g. checklists and standard operating procedures in the commercial world) just like many other safety advances.
Note that there are lots of planes flying with flaps switches next to landing gear switches, and flights in these planes have also experienced a reduction in gear related events by professional pilots. Private pilots, who do not have the professional’s training and playbook, still land gear up with too high frequency, much higher than professional pilots for commercial operators.
The fact is that in most cases, it is the “pilots” who messed up, and having a roadmap to better actions (checklists in aviation, better policy targets in monetary policy) would have increased safety/improved outcomes. Gear up landings aren’t a result of switch location, but usually by pilot error. Doesn’t matter where the switch is, if you have an unprofessional and unsystematic approach to compliance with procedures, you’ll get unforced errors.
Phil H
Aug 23 2021 at 10:19am
Thanks, that’s interesting!
Philo
Aug 23 2021 at 3:24pm
The lesson I draw from Phil H’s example is that the level of analysis matters for attributions of causation, which will figure in arguments about policy. Regarding a single accident landing an airplane (I’ll accept Jon’s correction, that the accident is retracting the landing gear when trying to retract the flaps), the low cost means of prevention is simply for the pilot to flip the right switch. But to deal with such accidents as a whole, the low cost “solution”—greatly reducing, though probably not completely eliminating, accidents of this type—may be to change the cockpit design, positioning the switches far apart. Bad design may then be said to “cause” the prevalence of such accidents, even though each individual accident is caused by pilot error.
Jon Murphy
Aug 23 2021 at 8:26am
It’s not clear to me how your story of the pilot is any different. It’s still a policy counterfactual: a “policy” of where the switch is placed. Institutional analysis requires policy counterfactuals. Indeed, in public policy, institutional analysis is policy counterfactuals.
Dylan
Aug 23 2021 at 8:58am
I think Phil is referring to the “lowest cost” portion of Scott’s post. In the pilot case, the lowest cost way of avoiding any individual accident would be for the pilot to pay more attention to which switch they were flipping. The way to avoid these kinds of accidents more generally is the higher cost redesign of the instrument panel.
I think that is a valuable insight when looking at policy more broadly (and one that I suspect you agree with). When looking at isolated cases of “bad” policy it can be tempting to figure out the lowest cost way of preventing whatever just happened from happening again and just doing that. But, then you’re fighting the last war and are probably not prepared for whatever comes next. A better way would be to look for structural issues that led to the problem the first time and try to fix those, even if they are higher cost in the short term, because they might be able to prevent a broader set of problems in the future.
Going back to Scott’s bus example. Let’s say that the steering on the bus had been installed backwards so that when the bus driver steered right, the bus turned left. Sure, the lowest cost way to avoid going off the cliff in that particular example would have been for the bus driver to steer left and the bus would have stayed on the road. But, we’d probably recommend fixing the steering on the bus if we want to prevent accidents in the future.
Jon Murphy
Aug 23 2021 at 9:53am
I agree that it is difficult to understand exactly Phil’s point since he never explicitly states it. That said, I’m not sure the “lowest cost” option rescues his comment. It’s not obvious to me that having pilots look for a second switch is lower cost than simply moving the switch, even if we assume a single case. Aircraft damage (not to mention time lost due to filling out accident paperwork) is extremely costly.
Alan Goldhammer
Aug 23 2021 at 7:16am
This is the same issue as with sudden acceleration in automobiles that first surfaced with the Audi 5000 way back in the mid-1980s (it still continues and Malcolm Gladwell had a podcast on this a couple of years ago. Audi spent a huge amount of money trying to correct this issue when it really was due to driver error. People did not want to believe they were stepping on the accelerator rather than the brake. In the WaPo story linked to, Brock Yates does the simple test of showing how sudden acceleration can’t happen.
Thomas Lee Hutcheson
Aug 24 2021 at 12:58pm
It is quite clear to me that monetary policy was wrong during 2008-2020; Inflation was too low when unemployment was too high.
I don’t think phrasing it as “monetary policy caused the great recession” is the best way to get the point across. It my sound like the argument that monetary policy was the cause of financial institutions failing to take systemic risk in their lending decisions and so was the cause of the September financial crisis.
Comments are closed.