The broader concept underlying "risk compensation"

Risk compensation - Wikipedia

Risk compensation is a theory which suggests that people typically adjust their behavior in response to perceived levels of risk, becoming more careful where they sense greater risk and less careful if they feel more protected. Although usually small in comparison to the fundamental benefits of safety interventions, it may result in a lower net benefit than expected or even higher risks.

I feel that risk compensation is just an example of a broader concept of “___ compensation”, in which things compensate for other things. Specifically, that you have negatives which have appeal but are held in check by being too negative, but if you do something to ameliorate the negative effect, that negative thing will naturally increase and cancel out the amelioration. An example is the notion that diet soft drinks cause people to consume more calories elsewhere.

Another example which involves different forces interacting would be something like municipal building policies. Suppose a town gets built up and this results in a big traffic problem, Now someone proposes that we solve that problem once and for all by building a whole bunch of new roads. Someone else argues that this will take the heat off the planning/zoning boards and allow them to be even more lenient about approving new developments, and before long you’ll be right back where you started.

Similarly, suppose a family has a tendency to overspend on credit cards and the budget generally. Is the solution to go all out to make more income? Not necessarily - that might just enable even more spending. And so on.

ISTM that this is fundamentally the same concept as “risk compensation” but not involving risky behavior specfically. Is there a name for the more general concept?

There are the following. But these are all names for similar processes in other specific contexts. I’m not aware of any overarching term that encompasses all of these.

Jevons paradox

Rebound effect

SnackWell effect

I can’t answer your actual question but there was recently a great article on Slate discussing that empirical evidence shows that risk compensation is somewhere between generally overestimated and completely non-existent.

People overestimating the effects of risk compensation has been one of my personal bugaboos for a while but I didn’t have any data. Perhaps I’m falling for confirmation bias here. I still need to read the underlying links.

I really don’t think much of that article. From the very first paragraph:

Remember March of 2020, before masks?..Among the complex reasons for the hesitation [to recommend masking] was a simple one: distrust of the public.

This is completely misleading. Risk compensation is not a matter of trust. It is a psychological process, often unconscious. I spend a lot of time hiking in remote backcountry in difficult terrain, and I carry an emergency satellite device that means if I break a leg a helicopter will probably come and get me within a few hours. I’m acutely aware that I have a tendency to engage in more dangerous behavior because I’m carrying this device. This has nothing to do with “trust”.

The reason for early concern about risk compensation with masking was because we didn’t know if masks would do any good, especially the crude masks that were all that were available to the general public. The idea that they “might not work but cannot do any harm” is just wrong. If they are not effective, then only a small risk compensation effect can negate the benefit.

This is an odd perspective when the whole point is that risk compensation is counterintuitive. You cannot trust your gut feeling, you need data that reflects unforeseen psychological effects in order to assess whether a proposed safety measure is effective.

Hmm. The diet example is self-licensing.

Past good deeds can liberate individuals to engage in behaviors that are immoral, unethical, or otherwise problematic, behaviors that they would otherwise avoid for fear of feeling or appearing immoral. We review research on this moral self-licensing effect in the domains of political correctness, prosocial behavior, and consumer choice. We also discuss remaining theoretical tensions in the literature: Do good deeds reframe bad deeds (moral credentials) or merely balance them out (moral credits)? When does past behavior liberate and when does it constrain? Is self-licensing primarily for others’ benefit (self-presentational) or is it also a way for people to reassure themselves that they are moral people? Finally, we propose avenues for future research that could begin to address these unanswered questions.

Merritt, A. C., Effron, D. A., & Monin, B. (2010). Moral self-licensing: When being good frees us to be bad. Social and Personality Psychology Compass, 4 (5), 344–357. https://doi.org/10.1111/j.1751-9004.2010.00263.x

(ETA: I see Riemann cited the specific form of self-licensing that applies to dieting.)

~Max

You could also be looking for “negative feedback loop”. Except these systems are all flawed; the equilibrium point is unfavorable. I don’t know if there’s a term for that.

~Max

It’s close to negative feedback, but that doesn’t seem to fully capture it. These are all describing phenomena where we change a primary input variable with a seemingly obvious causal relationship to the output, but the output feeds back to some other input variable, where the feedback to that other variable is in some sense unexpected or based on some unforeseen or subtle psychological effect.

The psychological effect would itself be the negative feedback loop, where the negative feedback kicks in later than we would like.

Insulin and glucose is the textbook negative feedback loop. So let’s make an analogy.

  1. As blood glucose rises over ~100 mg/dL, the pancreas secretes more insulin, which among other things lowers blood glucose by inducing glycogenesis. The input and output variable is your blood glucose level. The equilibrium point would be ~100 mg/dL.
  2. As blood glucose dips below ~100 mg/dL, the pancreas secretes glucagon, which among other things raises blood glucose by inducing glycogenolysis. The input and output variable is your blood glucose level. The equilibrium point would be ~100 mg/dL.
  3. As some hypothetical person who is bad with money spends over 150% of their budget limit, they panic, and cut spending down. The input and output variable here is spending as a percentage of the budget. The equilibrium point would be spending at 150% of the budget.

The difference is, spending 150% of your budget is unsustainable. The proposed solution from the OP, working for more income (thus having a higher budget) wouldn’t improve the situation because the feedback variable is spending as a percentage of the budget. With a higher budget the negative feedback (panicking) won’t kick in until even more is spent, which leads to the same problem again. This “solution” is in fact a positive feedback loop.

Compare #3 with a person who has type 1 diabetes mellitus, whose glucometer is defective and reads about 150 mg / dL lower than it should. It is still a negative feedback system, with the person administering the insulin injections replacing the pancreas; however, the equilibrium point is now ~150mg/dL higher than it should be.

A faulty “solution” comparable to earning more income above would be to somehow increase blood volume. But because the feedback variable here is glucose concentration, increasing total blood volume (if it were a thing) would not help. The negative feedback (insulin injections) wouldn’t kick in until blood sugar levels reached ~250 mg/dL, which creates a positive feedback loop.

~Max

Or in flowchart form,

Negative Feedback Loops

  Have you      --no--> Keep eating        Is traffic  --no--> Zone more stuff
exceeded your               |              unbearable               |
calorie budget? <-----------/                 yet?     <------------/
  |    ^                                    |    ^
 yes   |                                   yes   |
  |    |                                    |    |
  V    |                                    V    |
  Stop eating / eat less             Stop zoning stuff / zone less

Positive Feedback Loops

  Have you      --no--> Keep eating        Is traffic  --no--> Zone more stuff
exceeded your               |              unbearable               |
calorie budget? <-----------/                 yet?     <------------/
  |    ^                                    |    ^
 yes   |                                   yes   |
  |    |                                    |    |
  V    |                                    V    |
Increase the                               Build more
calorie budget                             roads

~Max

Risk compensation is counterintuitive. But the reaction among small government people is exactly as described in that article which is to greatly overestimate any reasonable effect of ris compensation. Whatever proposal is put forth to reduce people’s risk meets resistance that comes in the form of “if we try to do this more safely, people will just risk compensate to such a degree we’ll be worse off.” If you tell them not to drive drunk, they will claim they are a safer driver when they are drunk because they drive more carefully. Or joking-not-jokingly say taht the best safety feature in a car would be a big spike in the steering wheel pointed at the driver. Or if we give kids vaccines against STDs, they will just have more sex. And yet, when we add safety features to cars, passenger deaths drop. When we lowered drunk thresholds, drunk driving (and overall driving deaths) dropped. When we give kids STD vaccines, they get fewer STDs.

But none of that is comprehensive data specifically evaluating the amount of risk compensation. What is completely intuitive is that laissez faire, free market come hell or high water libertarians can justify their reluctance to endorse any government mandated public heath measures by latching onto thinly-supported by pseudo-intellectual arguments like risk compensation. I’m tired of their bullshit. .

I believe your specific Op is an example of a ‘wicked problem’:

Another related concept is the ‘unintended consequence’.

These are properties of complex systems, which are defined more by the interactions between things than by the things themselves, and which contain feedback loops, non-linear responses to inputs, and other features.