This seems conceptually related to risk homeostasis -- a perfectly self-evident yet oddly controversial topic (controversial I think because it makes a mockery of micromanagement attempts, particularly risk micromanagement by technocratic intervention).

The theory of risk homeostasis, which has been ably expounded by J Adams and G Wilde (separately), is that each invididual has an internally calibrated "riskostat" which is set to a comfort level idiosyncratic to that person.  So for example, if you tell that person that they driving a car equipped with superb automatic anti-lock braking, air bags, a secure seat belt system etc -- they will simply drive faster and more aggressively until their internal, subjective riskostat evens out again.

"Safety engineers" loathe this theory to bits, because they prefer a reductionist world where human actors are a static term in a mechanical model:  put hard hat on human, make human wear dayglo vest, make human wear seatbelt, and it's as simple as that -- instant injury/risk reduction.  However, in the real world all complex systems (including individual humans and societies/clans/neighbourhoods of humans) are interactive and have equilibrating behaviours;  they push back, as you might say, and adjust to new circumstances.  Critics of reductionist safety engineering sometimes refer to this as "consuming the safety margin".

Stats from the early days of seatbelt introduction -- when they were required only for drivers -- showed a sharp uptick in the number of injuries to passengers.  Drivers with seatbelts on felt safer and more secure, hence those with a medium to high riskostat setting consumed that benefit by driving faster and more aggressively;  those with a low riskostat setting benefitted by feeling less frightened and anxious while driving, i.e. they were lifted into their risk comfort zone rather than boosted out of it.

One of the best known RH studies is iirc the Berlin Taxi Driver Study, in which some percentage of drivers in a taxi fleet were told that their new taxis had state of the art ABS, and the rest were not told this.  (All the new taxis were in fact upgraded with ABS, to the best of my recollection.)  The taxis were equipped with GPS and accelerometers, so that speed and violence of manoeuvring could be recorded.  The upshot was not surprising;  the drivers who were told that their taxis were now "safer" and substantially improve, drove faster and braked a lot harder and later.

Another classic fingertrap of this kind is the "safe road" concept dear to traffic engineers, which means a road with extended sightlines (a euphemism for "no trees or other features anywhere near the roadway", very wide surface, very gentle graded curves, etc.  This road is supposed to be "safer to drive on" by presenting fewer surprises and obstacles to the progress of a motor vehicle.  However, in practise (and not surprisingly) motorists tend to "consume" these improvements by driving faster, until they reach their riskostat setting.  The only result of the road widening, tree removal, etc (aside from blighting the urban or rural environment around the roadway) is to set a de facto higher speed limit for the road, resulting in more risk to animals, pedestrians, cyclists, etc.

There's a suggestion in the literature that after basic needs are met (no one is very happy if they are involuntarily cold, hungry, or at physical risk), individuals may have personal "felistats" which recalibrate to a equilibrial level of happiness;  and that absolute material wealth is far less important than positional perception.  A chunk of evidence -- very intriguing stuff -- suggests that perceived inequality, i.e. having one's nose rubbed in the much greater wealth and privilege of other people, has a significant effect on general physical health as well as mental attitude;  that perceiving others as "ahead" or "better off" all the time produces in a large sample of people a generally elevated stress response, with attendant insomnia, indigestion, depressed immune system function and so on.

All fascinating stuff, but I think (OK, I have a bee in my bonnet on this, I admit it) the salient lesson in each case is that complex systems are self-equilibriating and respond to single-factor tweaks in ways that often defeat the tweakers' intent.  The human ability to adapt scares the bejeezus out of me, to be quite honest.  It is already "normal" for kids growing up in southern California that more than half the beaches they know of are closed for health/safety reasons due to urban effluent.  That's just how things are, and they're used to it.

This is the problem of the "shifting baseline" which has perplexed and saddened many a social justice or environmental activist;  that we adapt and become inured very quickly to conditions which only a generation (or less) earlier would have stirred outrage and a demand for action.  Almost anything can become normal, ordinary, background noise -- already many people are accustomed to the idea that e.g. the US has occupied Iraq and "there's a war going on" and why not?  That's just how it is.  The (deliberate?) shortness of attention span of US media, the cultivated amnesia and decontextualising of information, surely contribute to this problem...

Hmmm, I think I'm rambling.  But by now everyone is accustomed to that, right?  ;-)

The difference between theory and practise in practise ...