by Pierre Lemieux
Philosophers have been debating moral values for two and a half millennia, and counting… I would argue that, at least when moral philosophy becomes political philosophy, consequences matter.
According to FBI statistics, 39% of all murders (counting only those where the murderer’s age is known) are committed by young males aged 17 to 24. These murders impose great costs on the victims and, perhaps even more, on their parents, children, and other loved ones. The internment of all young males from their 17th birthday until they turn 25 would prevent these murders. Therefore, this public policy should be implemented.
Many people seem to believe that public policy should be based on this sort of reasoning, which can be expressed under the form of a syllogism:
X imposes great costs on some people;
Z would reduce the occurrence of X;
Therefore, Z must be done.
Call this the simplistic public-policy syllogism. (I was reminded of this sort of public-policy model by a Facebook comment from independent researcher Daniel Kian Mc Kiernan, which he later developed into a blog post.)
As a preliminary consideration, we want to make sure that the minor premise (the second one) of the syllogism guarantees a net reduction of X (murders committed). Z (the preventive incarceration of all young men) would not eliminate the roughly 6,000 murders committed annually by young men, for some would be committed in their prison–especially with all these testosterone-laden young guys around. But it would reduce the occurrence of murders. It would save at least some lives (if not most). As public health activists are fond of saying, “If it could save only one life…”
Even if the two premises are correct, the syllogism is logically invalid. The conclusion does not logically follow because, as philosophers know, an “ought” cannot be derived from an “is.” There is a jump between the positive and the normative. A normative conclusion requires the input of moral values (often called “value judgments” by economists).
Philosophers have been debating moral values for two and a half millennia, and counting. The deontological version of ethics is usually based on some conception of natural law. I would argue that, at least when moral philosophy becomes political philosophy, consequences matter. A deontological ethics that would render everybody on earth miserable would be indefensible. Consequences, however, must still be evaluated.
We need to evaluate the consequences of a public policy compared with its absence (or with its absence in the presence of a different public policy, but I ignore this complication here). Assume that we are only interested in individuals–as opposed to, say, what Gaia thinks or what “society” feels. All the costs and benefits of Z to all individuals must be factored in. The concept of opportunity cost is useful here (as everywhere): benefits foregone are costs, and costs avoided are benefits. Thus, maximizing benefits is the same as minimizing costs over the two alternatives Z and non-Z. And, needless to say, all psychic costs and benefits must be included.
The cost of Z would include the pleasure lost by the young men during their preventive incarceration. Foregone schooling would translate into reduced future earnings and enjoyment of life. The cost of Z also includes the benefits lost by other individuals as a result of the internment policy: parents deprived of their sons during seven years, young women deprived of young men, the loss to taxpayers who have to pay for the jails and guards, and so forth.
Standard welfare economics and its practical application, cost-benefit analysis, are not capable of eliminating the need for moral criteria in evaluating the consequences of public policy. Even if the total benefits of Z are greater than its total costs, some individuals will be harmed. The idea that the individuals benefited by Z could potentially compensate those harmed, even if they actually did it, is not sufficient, for what allows government to favor some individuals over other individuals? This is a standard and well-taken critique of the New Welfare Economics that developed in the 1930s.
Economists have demonstrated that, if everybody is to count equally, it is impossible to aggregate all individual preferences into a sort of “social welfare function” representing a moral agreement about how some can be harmed in order to favor others. I present an introduction to these proofs in my Econlib article “The Vacuity of the Political We.” Somebody–individual, group, or majority– will typically impose his value judgments on all others in society.
Add to this a crucial fact: long-term consequences that must be factored in are impossible to predict. How many young males would have become hardened criminals when they are released from their prisons, and will they commit more murders than they would otherwise have? What would be the consequences of this assault on individual liberty in general? How will all this change the nature of society and the state? The character of the jailers? The relations between boys and their families? (Would parents start aborting their male foetuses as many Chinese did to their daughters-to-be because of another sort of public policy?) Marriage and the family? Conventional morals? And so on and so forth.
To summarize, the simplistic public-policy syllogism is useless and dangerous. It is not logically valid. The fact that X imposes great costs on some people is not a sufficient reason to implement Z, because all costs must be considered. Accepting a distribution of these costs and benefits among different people requires value judgments, even if the net benefit is positive. These value judgments will have to be imposed by a majority (at best) to the rest of society. Moreover, long-run costs and benefits are impossible to forecast.
There are ways to circumvent those problems and create a (narrow) space for public policy. We may invoke James Buchanan’s social contract, Friedrich Hayek’s free society principles and spontaneous rules, or perhaps even Anthony de Jasay’s (modified) presumption of liberty. But none of these approaches justifies the simplistic public-policy syllogism.
READER COMMENTS
dave barnes
Feb 24 2018 at 11:53am
Or, we could conscript every 17-year-old male into the army for 8 years.
1. Less likely to commit murder.
2. We need the cannon fodder for our current wars and the upcoming ware with China.
kingstu
Feb 24 2018 at 12:29pm
A Few Islamic Terrorists impose great costs on some people; Incarcerating All Muslims would reduce the occurrence of Muslim Terror; Therefore, Incarcerating All Muslims must be done.
A Few Gun Users impose great costs on some people; Eliminating All Guns would reduce the occurrence of Gun Violence; Therefore, Eliminating All Guns must be done.
If Liberals won’t incarcerate all Muslims due to the actions of a small minority of Muslims, why do Liberals want to eliminate all guns due to the actions of a small minority of Gun Users?
Same question can be asked to Conservatives by switching Muslims and Gun Users.
James
Feb 24 2018 at 1:55pm
Pierre:
Everything you say is true but I doubt anyone would be persuaded by it even if they agreed with it. If you could get them to be perfectly honest, they would likely say “I know my preferred policy is based on flawed reasoning and imposes my values on others. That’s not important. Implementing the policies I want is important.”
Kingstu:
I know that some liberal politicians do want to ban private firearm ownership but I’ve never heard of any conservative politician calling for the internment of all Muslims or an outright ban on Islam.
Pierre Lemieux
Feb 24 2018 at 3:58pm
@James: Your first point is a good point. But aren’t intellectual arguments aimed at bringing people to be honest and understand the real reasons for their arguments? The next step is something like this: “You want to impose Z, which would greatly harm me; I would like to impose W, which would greatly harm you. So let’s agree to have individual liberty instead.”
Daniel Kian Mc Kiernan
Feb 24 2018 at 7:00pm
Discussions of consequentialism versus deontologism tend to be miscast. When definitions are discussed carefully, some people who have imagined themselves as consequentialist will conclude that they are deontologists after all, and vice versa; and some people will conclude that there is a non-empty intersection, perhaps even one that they occupy.
A very naive consequentialism assumes that we can choose policies, each of which leads to some end-state, and that we should choose that policy which leads to an optimal end-state. A more sophisticated consequentialism recognizes that outcomes are imperfectly predictable, and seeks one of those policies that have optimal valuations for their sets of possible end-states, given the relative plausibilities of those end-states under each policy. It has been noted that consequentialism is almost invariably specifically associated with using an expectation and with maximizing that expectation. Someone concerned about consequences but not attempting to maximize an expectation might seek a different self-identification.
Deontologism insists that there is some policy that should always be chosen, regardless of actual outcome. But the deontologist may be speaking of a realized outcome or of the most plausible outcome or of the valuation of the set of possible expectations. (Some clever readers will see something here which will be noted below.)
Consequentialism tends to be situational, choosing the policy that is believed will lead to the best outcome or has the optimal expectation in the present situation; deontologist tends to choose one policy across all situations. This distinction, however, dissolves if we recognize that all policies can be conditioned, that a universal policy filled with conditions will operationalize like distinct policies for distinct situations, and that in reshaping policies to fit situations, people are generally guided by underlying principles that can and ought themselves to be rational.
The real question, as I think, proves to be that of how the past shall bear upon the the present.
Say that we could view absolutely everything about two possible states of the world, but could not infer a path by how those state could be achieved; each were associated with a black-box policy. We might compare the two states, and decide that one were better than the other. Could we choose the black-box policy from this information? If we learn that what seemed the better state were achieved by repulsive means, has the better state changed? Or was the past always there but simply hidden (within the black box)?
If we include the past _in_ the present, then our valuations of the sets of possible outcomes of a policy are determined to some extent by that past, by the policy itself, and we might reject a policy without knowing anything more about its set of outcomes than that they would be achieved by such a policy. (This is what the clever reader saw earlier.) I believe that this is the principle truth for which the deontologists have been grasping. The means are amongst the ends.
(As an exercise, consider the implications for the strong independence axiom.)
Adam
Feb 25 2018 at 11:26am
Excellent point–so important and all to easy to ignore.
andy
Feb 25 2018 at 6:42pm
I actually tried to counter this argument – I asked if the opposite number could give me some argument, why he thinks that proposed policy’s benefits are higher than the costs.
I always get answers that surprise me. This time the answer was essentially that it doesn’t matter and that the benefits are incalculable.
Next time I will ask if they are incalculably high or low… and if it really doesn’t matter if the costs are higher than the benefits… I’m always surprised what people claim when they support ‘their’ argument. I’m slowly getting better at answering these extreme claims, but I’m still getting suprised.
Comments are closed.