Economists like to talk about the 'efficiency rates' of various things.

For instance, it would be possible to design and build a car which would last for 100 years, but one one would buy it. It's performance would be much lower than the cars we see today which last for 20, and may break down dozens of times in that 20. The imperfect solution gives us the better result. There is an efficient level of vehicle break-downs.

It is possible to evolve an animal with bones that would never break - but such an animal would be heavier and slower than their naturally evolved counterparts, and would by any reasonable measure be inferior. There is an efficient level of bone breakages.

There is, they tell us, an efficient rate of burglaries. This is the point at which the rate of burglaries is optimal for the tools available to control it.

If crime prevention measures are decreased from this point (which might save resources) - the burglaries increase more than than the money saved. If crime prevention measures are increased (which will cost resources), the savings from the reduction in burglaries won't be worth the extra cost.

There are two main criticisms of this model. The first criticism is that many of the costs (and benefits) are hidden or incalculable, but this doesn't change the underlying principle - the costs and benefits are real, even if they are hard to calculate. Putting a cost on being burgled is hard, but it is clearly a real figure. Few people would be willing to give up their DVD and some of their dignity for a dollar, but most would do so for a million dollars - so the cost of their loss is clearly between these.

The second criticism is that if only we tried formula X, the problem (or part of it) would just go away. Formula X is usually some weird and unworkable solution, like abolishing private ownership, having group love-ins, or employing more social workers. Sometimes formula X is quite sensible and workable - in which case it becomes part of the available technology, which when implemented, and becomes part of the efficiency formula.

This does not just apply to economic bads - but also to all economic goods. There are efficient levels of timetable delays, equipment failure, software errors, car crashes, jaywalking, divorce, shoplifting, stock-market crashes, rape, child abuse and murder. The real world makes compromises.

And there is an efficient rate of all economic bads - pollution, robbery, muggings, rape, torture and murder.

Technological improvement may lower the efficiency rate for the bads, but is unlikely to reduce it to zero.

Any claim that perfection somehow takes this into account is really just changing the definition of perfection to mean efficiency - a plausible but not intuitive definition.

Someone who strives for perfection is not striving for any greater good. They are not utilitarians.