I've certainly heard all the various theories one way or another, about the only one that really holds solid traction for me is that a non-homogenous culture has a much wider space between law and taboo, and therefore more crime. Regardless, while I don't like a lot of our laws and don't entirely approve of how we rehab or imprison people, the moral point remains, we must not change our laws for the purpose of decreasing the prisoner population. Laws should exist because they are moral, not convenient. "There'd be a lot less people in prison if X was legal" is an argument which holds no water with me. I think pot should be legal, maybe other drugs too, the effect on the number of prisoners or the advantages of being able to tax it are, to me, BS arguments.

Whatrever the case, morally, it is still possible to compare the economics of the issue. The same libertarian in me that say "If they want to snort coke, fine, they can drink cyanide for all I care" does not translate to saying "Hey, let's rip the safety warning off stuff, if you're dumb enough to want to make toast in the shower, good riddance"

All resources are limited, and while the net resources, effectively the GDP for most practical purpose, tends to grow over time both as a total and per capita, at any given time it's basically a zero sum game. If we have X amount of money to spend on firefighters and cops, and 40/60 split appears to save more lives and cause less damage than a 60/40 or 50/50, you go that route. In most cases where it doesn't stomp on liberty too much, you act pragmatically, if that means you find the best path is to to send everyone a fire detector for their house, complete with $100 tax rebate for anyone who sends in a picture proving they installed it, then ship them a 9V battery for it every year with a letter reminding them how important it is, then that's a legitimate and cheaper expenditure than a large fire department. Maybe it isn't. Maybe doing that and having a large fire department is better than having 10 parks in the region when 5 instead wil make up the difference in revenue.

Now, we really do this, we do it pretty well in democracies and even in a lot of the more outmoded systems it was in effect, but emotion, special interest, bribery (usually subtle), these things tend to make people look at a case where X gets 50 cents on your Euro in return to Y that gets the whole Euro and think X is the better deal. In some ways, the intellectual themselves is the problem. It is far too easy for those of us who have whole walls dedicated to books to say "You can't cut the hours on the library, that's mad!" when in reality it tends to be an area of some waste. Ditto public schooling, we want people to know so much that we, right or wrong, simply will not accept that a lot of money spent on higher education is wasted not just in general inefficiency but by bothering at all. When we spend $10,000 or whatever to educate 20 people on the basics of biology, have we done them as much good as if we had spent that money on, say, vaccines?

A million for defense but not a penny in tribute, sure, morality sometimes has to trump pragamatism, but when will a million in foriegn aid save ten million in military expenditures, or $1.00 dollar for a sniper's bullet circumvent both?

How often do we double the expenditure for one service only to basically get a 50% or so increase in result? Is it worth it? Maybe, I know the rationale we use, that kid who didn't have a library and didn't have a biology class also didn't invent the cure for AIDS. But maybe he did have those then got shot on the way to school because we didn't have enough police, or died of injuries because the taxes were so high he didn't maintain his car properly and got in an accident, and the money that paid for his education didn't pay for our hospitals and the doctors didn't save him on the operating table. It's a balance thing, and where prisons are concerned, while I agree we have too many people in jail, I more tend to think we just don't turn them right... starting from the core assumption that sticking people away from potential victims by sticking them with a lot of other criminals. This, to me, has always seemed about as logical as trying to cure sex addicts by sticking them all in a room together. It's also a better case for decentralization, because in that model you get to see more variety in systems and people don't have to fight as hard to try out new ideas, and goods ideas when tried out tend to win. IT is far harder to see this when things are centralized.