Compensatory concealment: or why good things are generally better than they appear to be and bad things are generally worse than they appear to be.
I want to introduce the idea of compensatory concealment because it’s essential for understanding almost all changes involving people, but I’ve never heard it talked about explicitly. Like the distinction between the effect at the margin and the average effect, it’s an absolutely crucial idea, imported from economics, that most people aren’t aware of. However, unlike margin versus average thinking, I’d say that even many economists don’t think clearly enough about the phenomenon, at least outside obviously economic contexts.
There are applications in fields ranging from economics, epidemiology, psychology, environmental management, political science, traffic engineering, sociology, anthropology, history, religious studies and just grokking people on an everyday level. Simply: I want to add it to the mental toolkit of everyone.
Compensatory concealment means that most apparently good changes are better than they appear, and most apparently bad changes are worse than they appear. It is perhaps best introduced through an example.
The death rate on Autovania’s roads is Y. Let’s suppose that, tomorrow, every Autovanian miraculously becomes a better driver. For a little while, the death rate would fall to some level X. Then, slowly, the death rate would rise again to some level X’, lower than the original Y, but still higher than X.
Why is this? A couple of reasons. Most directly, with drivers being more skilled, people would relax a bit on the roads. Over years, the auto industry standards might be relaxed, or at least the rate at which their severity increased might be reduced. Traffic engineers would concentrate more on speed and less on safety.
So we have a visible good change- a decrease in road accidents, and we have an invisible good change, a reduction in effort and stress devoted to avoiding crashes. The existence of the invisible good change means that there isn’t a full “flow-through” of the positive effect into the primary variable of concern (crashes). Whether we wish there was more flow through or not is a judgment we must all make ourselves, but a reduction in stress and effort needed to stop crashes is undeniably a good thing in itself.
But this turns out to be a highly general phenomenon, true much more often than not. When something happens to improve a variable we care about Q, Q often (usually?) does not gain the full benefit. This is because other compensatory mechanisms designed to prop up Q if it’s good or keep Q down if it’s bad, might be removed or reduced in scope, as those who judge their cost effectiveness no longer think that they’re worth it.
All these facts have a number of consequences:
As alluded to above in the Autovania example, good changes tend to be even better than they appear.
Bad changes tend to be worse than they appear. An example of this: a cheaper medicine treating condition X goes off the market, and only an expensive one is left. We see less treatment of condition X, and poorer outcomes as a result, but things are actually worse than this. Even though the total number of doses of medicine to treat condition X will fall, the amount spent in total on doses will likely rise, and this will lead to a less obvious negative effect- less treatment by the health system for other conditions.
State policy, across the board, is more powerful than it looks in both its good and bad effects.
There’s some economic language for a related situation. Suppose that rice gets more expensive in a society where most food is rice. On one hand, this will tend to make people choose alternative foodstuffs to rice because they’re relatively cheaper. This is called the substitution effect- People substitute other foods with rice because rice is now relatively expensive- at least compared to what it used to be.
On the other hand, this will tend to make people tend to buy fewer alternative goods because all their money is being used up on that damn rice. This is called the income effect- people have less remaining income because the rice is more expensive. Generally speaking, in cases like this, the substitution effect is stronger than the income effect, but in theory, the income effect can be stronger, creating a Giffen Good.
Go back to our case of Autovania. Because units of safety cost fewer ‘resources’, people have more resources left over to spend on other things. It’s quite likely that some of the resources freed up by easier car safety will be put to other uses. This is basically a form of what economists call the income effect. However, since we’re dealing with resource expenditure- a broader notion than ‘income’, we’ll call it the resources effect to signify it’s not just about money.
Yet because being safer is easier, people will choose to get more units of safety. This is a form of what economists generally call the substitution effect. That name will work for us too.
What about the case where the change is bad rather than good? I’ll leave that as an exercise to the reader but know that it still involves a resource effect and a substitution effect. In both cases though- negative and positive-, we only directly see the benefit (or harm) represented by the substitution effect. As the good thing becomes easier to obtain/harder to obtain, we buy more/less of it, but our increased/decreased purchases of other things with our “resources” don’t appear in the tabulation.
In some ways, this is reminiscent of Bastiat’s discussion of the difference between that which is seen and unseen, much as I loathe the political uses and naïve way people talk about that.
Compensatory concealment should always be considered as a possibility when looking at the good and bad effects of nearly everything:
Some hypothetical examples
Some positive cases:
An SSRI reduces anxiety. With the new energy people have, they go do more things, this increases their anxiety somewhat, though not to the prior level. The patients experience part of the benefit of the drug as reduced anxiety, and part of the benefit as increased activity.
A new drug increases metabolism. It makes people lose weight. As a result, they feel like they can eat a little more. This increases their weight somewhat. The overweight experience part of the benefit as reduced weight, and part of the benefit as having to worry less about what they eat and enjoy more foods.
A new initiative reduces burglaries. As a result, people feel like they can leave their doors unlocked at night, which somewhat increases the level of burglaries. The home-dwellers experience part of the benefit as being less likely to be robbed, and part of the benefit as having to take fewer security measures.
Now for some negative cases:
A spike in temperatures due to global warming causes indoor temperatures at work sites to increase, however a graph of daytime worksite temperatures wouldn’t show the full harm, because those temperatures are being partly offset with costly, power guzzling air-conditioning.
An economic crisis reduces wages. Some workers don’t have their wages reduced as much as they might be but accept other changes- less paid leave, intensified pace of work(*) etc.
Quick note
Compensatory concealment is only one kind of unintended consequence. There are plenty of others that don’t fit this pattern. A single action can cause multiple types of compensatory concealment, sometimes both good and bad, etc.
Footnote:
(*)The economist John Quiggin makes a great point about intensified pace of work in the decades after 1970 as a source of economic growth. Economists often present it as a win-win. It is absolutely not a win-win. It has a real cost from the worker’s perspective- a more stressful work environment, less capacity to engage in leisure activities at work, less capacity to build social connections at work, etc.
Some people respond to this by saying that they don’t care about these costs because they feel the boss “deserves” the worker’s full capacity and attention at work, thus any losses by workers are of something they are not entitled to. This is an insane view. The minimum wage coffee maker shouldn’t be expected to use up every ounce of their energy, working like a trauma surgeon trying to save a child, for 7.25 an hour. The idea that you should “give a 100% at work” is either crazy or is coming from people who never thought very hard about what 100% means. Workers deserve a bit of space for a chit-chat and a phone scroll.
If you enjoyed this, subscribe for free to my Substack. Also, come read my book
I wrote a book. Get a copy:
https://www.amazon.com.au/Cave-Scratchings-philosophical-left-wing-miscellaneous-ebook/dp/B0B8XXSKJQ
Bonus: Some basic thoughts on the Repugnant conclusion that don’t merit their own essay
A point that many commentators have made is that life barely worth living is, definitionally, worth living.
However, another point is that when you are assessing the repugnant conclusion *you can choose for yourself what the threshold is for a life worth living*. It's your values system we're assessing the consistency of. You can set it on whatever level you like. If you think people are deluded into thinking their lives are worth living, when actually they are underneath the threshold, then you can factor that in. When you think about a life barely worth living, if you shudder or weep for them, then you're not really thinking about a life barely worth living.
You can also set the currency in which the balance of the value of a life is judged, whether it's happiness, desire satisfaction, or my preferred option- Eudaimonia- a complex of features including creativity, learning, happiness, desire satisfaction, social connection, meaningful autonomy, etc.
Imagine a billion blissful people living wonderfully meaningful lives. Now imagine trillions of trillions of people whose lives are filled with great joy and sadness, wonder and drudgery, hope and fear, freedom and compulsion, virtue and vice. Complex, beautiful, ugly, flawed, suffering, joyous things, and their lives are worth living- they would affirm it all and do it again, even if only by a hair's breadth. Now, I think, the repugnant conclusion doesn't look so repugnant.
Even if you would hesitate to move away from the blissful billions world to the 'repugnant' trillions of trillions world, I think it's also true that you would hesitate to move away from the repugnant trillions of trillions world to the blissful billions world. Thus any remaining intuition has more to do with a preference for the status quo than any supposed repugnancy of the conclusion.
Interesting and useful! Seems related to idea in ssc/acx post about 'straight lines drive reality'. https://slatestarcodex.com/2019/03/13/does-reality-drive-straight-lines-on-graphs-or-do-straight-lines-on-graphs-drive-reality/
Great concept. The "regression to mean" phenomenon that seems everywhere isn't just statistical or selection bias (we notice stark changes, but stark changes are outliers), but instead, they are natural to the multi-dimensional web of incentives in which everything exists.
Anecdotally, I joke about how when I started taking hangover pills, they worked until I started drinking twice as much.
Or: We've done lots of work on life-extension over the last hundred years, but we've compensated for it with life-shorteners, such as skydiving and drug abuse.