To start with, the fundamental concept of Pareto Improvement has to be explained.
A) Imagine a situation of two people A and B, trading with each other. IF
The trade can result in a position that both A and B are better-off (or feel happier),
Then the trade results in a Pareto Improvement. This is the simplest situation.
* * * * *
B) Now, imagine A is doing some damge to B (e.g. A kicks B's ass). IF
1) B has the legal right to have his ass unkicked AND,
2) A by paying B some money (or other kinds of economic benefit, like a kiss, or a one-night-stand) can get B to agree that A can kick B's ass AND,
3) A is willing to make such an offer so as to kick B's ass,
Then there will be a Pareto Improvement when A makes an offer to kick B's ass and that B voluntarily agrees to the offer. Because everyone will be happier.
* * * * *
C) Now, imagine A is still kicking B's ass, IF
1) government knows perfectly what A and B want (economically, their marginal valuations),
2) government orders A to compensate B for kicking B's ass,
Because government knows everything, so the ultimate result will be exactly the same as B) and everyone is still happier. So it is still a Pareto Improvement.
Therefore, we come to a rather unconvential and unpopular economic conclusion: as long as party A can compensate party B in a way that party B will agree, party A can do anything on party B and it will still result in a Pareto Improvement (i.e. everyone is happier than before). And by anything, I strictly include anything, like kicking ass, slapping, molesting or even raping EXCEPT (by now) killing the other party.
(This kind of conclusion partly explains why economics is not a subject that can be understood by all --- it gives inhumane, immoral conclusion. BUT to clarify, the analysis is strictly and only based on economics AND it is not suggested that we should not consider anything else before applying such conlusion, no economist would have made such a claim that only economic principles should be followed while making decisions)
Simple and sensible, it is just not possible to compensate the other party satisfactorily if the other party is dead.
On the other hand, government can (economically) also order anyone to do anything (that derives a benefit) if at the end of the day the derived benefit can be re-distributed in a way that everyone is happier since there will still exist a Pareto Improvement.
But again when anyone is dead, there is no way to re-distribute anyting to him/her anymore and not strictly everyone will be happier and it will not be a Pareto Improvement.
There is, however, one empirically rare exception. It is when party A values party C's benefit so much that, even by re-distributing the derived benefit to party C, party A would have felt happier (well... in heaven/hell) even suffering the loss of life. Such a case only happens when someone voluntarily sacrifices his/her life for someone else's life (Titanic, maybe?). But the emotional impact from the Titanic movie somehow sends us a message that this REALLY does not happen so often, if at all.
Thus, it comes to a conclusion that
I) nobody should have sacrificed their lives for anything (but economically acceptable if stricly voluntary),
II) no government should have ordered their citizens to sacrifice themselves for anything because re-distribution is close to impossible as the exceptional case proves to be really rare empirically.
( conclusion II should be more strictly followed than conclusion I, because in I, individual makes his/her own decision and it can be comfortably presumed that any result is voluntary, not so much in case II)
So war for high-moral-ground values, slower poverty reduction (which kills) to preserve cultural heritage? No way as long as the Titanic kind of movie can still touch everybody's heart.