In the previous few posts I discussed two of the more interesting ways ethical or value or normative propositions not based on “utility” are informally added to neoclassical welfare economic theory. I described those rhetorical constructions as belonging to the most ethically controversial outer layer of the metaphorical onion of bad economics. That’s the layer that deals with ideas and sentiments that seem to me the most distinct from the core elements of that theory and, in fact, aren’t really part of that theory at all but tend to get associated with it in practice. This week I thought I’d go back to the basics and discuss the explicit or implicit ethical proposition we express when we say or imply we should maximize total social utility in neoclassical welfare economics. Drawing out the logical implications of that proposition is really the whole point of neoclassical economic theory. If one isn’t interested in that proposition, one should just not bother reading that particular economic theory at all. This is the bit I described previously as the core of the onion of bad economics because if we don’t have this proposition, then we don’t really have the neoclassical welfare economics on which bad economics is based. Fair warning, this one is a long one. Didn’t necessarily plan it that way, but that’s the way it worked out.
The first thing we must address is the inevitable philosophical baggage that goes along with the word “utility.” Fortunately, I did that in a previous post, Preference Utility and Fake Utilitarianism (April 15, 2020), so hopefully I won’t have to go over that again here and I can just dive straight into what the ethical propositions involved actually are rather than explaining yet again what they are not. Just a quick reminder, the crucial element of the concept of “utility” as defined and used in neoclassical welfare economics is that it cannot support interpersonal utility comparisons, which means it cannot be used to resolve interpersonal conflicts. What we’re really expressing in the proposition we should maximize total social utility is we should allow other people to do what they prefer to do or want to do if there is no interpersonal conflict involved. The concept of “utility,” per se, is superfluous and serves no role other than to act as a red herring. That’s the case under both common interpretations of utility one finds in discussions of neoclassical economic theory: “preference utility,” the easy case where utility doesn’t actually exist at all, and what I call “perception utility” where utility exists, in a weird sort of internal perception sort of way, but does not represent anything anyone would sincerely be interested in wanting to maximize if he or she could.
Seems simple enough but, of course, not really. Nothing ever is. The world can be quite annoying that way. To think about the implications of maximizing total social utility without the additional and unrelated normative or ethical content associated with accepting property rights and using market mechanisms to resolve interpersonal conflicts based on economic power, let’s consider a world with no property rights, no money, no markets, no market transactions, no compensating people for externalities, and so on. Let’s just think about the proposition we should maximize total social utility in isolation. Don’t worry, I’ll discuss what happens when we add the unrelated propositions about property rights and markets when I get to the normative propositions introduced in what I’m calling the middle layer of the onion of bad economics, which concern normative or ethical propositions that appear to be part of neoclassical economic theory but are not based on utility and not even expressed using the word “utility,” because of course one can derive rather different results from combinations of ethical proposition than just the one.
So let’s break it down. First of all, we need to distinguish between the normative proposition of interest here, the fundamental ethical proposition on which neoclassical welfare economics is based, that we should maximize total social utility, from the tautological statement one finds in economic theory that people act so as to maximize their own personal utility, which basically translates in plain English to people preferring what they prefer. The ethical or normative proposition on which neoclassical welfare economics is based addresses what one should do with respect to other people. It’s basically a form of social ethics. So the first thing to realize is that to analyze this proposition we need at least two people to be involved in some way, the person reading neoclassical welfare economics and trying to evaluate the ethical proposition about maximizing social utility, and another person, who can give the first person some context in which to apply the ethical proposition in question.
After a bit of head scratching, it occurred to me there are actually two ways to think about what’s going on here. One approach, the approach I think is the easiest and most intuitive and the one I’m taking here and also took in my most recent book on the subject, is to imagine the person reading economic theory and thinking about the normative proposition we should maximize total social utility looking in from the outside as it were, which allows us to imagine the observer applying it to someone else in a one person world, which has the advantage of allowing us to assess the proposition in the absence of interpersonal conflict. An alternative approach is to place the person reading economic theory and thinking about the normative proposition we should maximize total social utility in the world along with the person acting as the subject. The complication in that case is that the observer becomes a part of the society he or she is studying, which means the assessment of the normative proposition we should maximize total social utility by the observer shows up in practical terms as an introspective determination of his or her own ethical beliefs about the behavior of the other subject and hence a determination of interpersonal ethical conflict with the other subject that cannot be resolved on the basis of utility, which in turn makes the proposition we should maximize total social utility empty, irrelevant, inapplicable, devoid of practical implications. I think the two ways of looking at this issue are comparable and the difference is really only a matter of how one chooses to talk about it. When ethical objections to the proposition arise the resolution in the one case is to reject the proposition and act presumably on the basis of some other ethical proposition, and in the other case to discover the proposition is empty and act presumably on the basis of some other ethical proposition. To demonstrate the equivalence I may try to write up the following bit of this post in the second style in a future post. Really, I suppose the choice depends on how comfortable one is evaluating ethical propositions relating to how one should treat another person in a one person world.
For now, say we have two people A and B. A is reading neoclassical welfare economics from a great height and wants to think about the implications of maximizing total social utility in terms of how he or she should act toward B were he or she to descend the great staircase and start interacting with B, who currently is on his or her own in a sort of one person world down in the valley somewhere. What is the implication of A trying to maximize total social utility in this context? What does it really mean? It means A should allow B to follow his or her preferences, whatever those may be, since by the way we’ve constructed the scenario none of those preferences could generate an interpersonal conflict that would be unresolvable if we confined ourselves to thinking about utility as it is defined in economic theory. So let’s imagine A evaluating the ethical proposition we should maximize total social utility.
The easiest case would be where B is considering doing something ostensibly innocuous like choosing between eating an apple or an orange. B determines he or she prefers to eat the apple, and A imagines himself or herself maximizing total social utility by saying, sure, I’m not interesting in stopping B from eating the apple, that is, in interfering with B’s attempt to express his or her preferences in the absence of any interpersonal conflict. Why would A think otherwise? Of course, A is not a simple minded fool. He or she knows an accepted principle of ethical philosophy is that one doesn’t just look at the simple case and move on. One tests one’s ethical propositions by trying to think of difficult or tricky situations. So let’s do that.
Let’s say A realizes B does not have full information about the choices B is making and further A knows something B does not. Let’s say, for example. A knows the apple in question has been infected with a deadly pathogen, but B does not. B prefers to eat the apple, unaware it will likely kill him or her. What does A think about the ethical proposition he or she should maximize total social utility in this case? Seems like that one might be a bit more ethically controversial. A may have to think about that one a bit more. He or she may indeed still support the proposition we should maximize total social utility saying something like, I’m not B’s nanny, or if B can’t be bothered to determine if the apple contains pathogens it’s not my job to explain it. On the other hand, A might feel he or she has an ethical responsibility to intervene to help B better understand what he or she is doing. In that case, A would reject the ethical proposition he or she should maximize total social utility in this case. A could, of course, maintain support for that proposition by introducing some sort of stipulation or condition that would preclude this issue, so something like, we should maximize total social utility as long as B has full information about what he or she is doing, or since full information seems a practically difficult and costly prospect, maybe something a little more modest like a reasonable level of information, or maybe something like as long as A provides B with any relevant information A might have.
Interestingly, this discussion of the potential relevance of information and information asymmetries for evaluating the ethical proposition we should maximize total social utility may bring to mind to students of economics anyway the discussion of full information or perfect knowledge as a condition for an ostensibly socially optimal perfectly competitive market. It’s clearly relevant there as well, albeit in a somewhat different way, and I’ll discuss it again when we get to that middle layer of the bad onion of economics. However, the role of information is clearly rather more foundational than defining perfectly competitive markets. One thing I think worth mentioning here is that in the context of defining a perfectly competitive market the assumption of full information is often introduced in the context of a scientific theory or model, and in that context the assumption of full information can famously be treated as what is known in science as a simplifying assumption, a pseudo-factual assumption one understands may not be true but is instrumental in a theory that may nonetheless be highly evaluated as a whole on a scientific basis, that is, in terms of its ability to predict or explain observable phenomena. That is not the context in which we’re raising the issue here. We’re talking about evaluating an ethical argument in which this issue of information appears as a factual premise. In the context of an ethical or normative argument, whether or not the factual premises in the argument are true or not is crucial to the evaluation of the argument. If this issue is relevant to our ethical argument, and we’re meant to take the proposition we should maximize total social utility seriously, then we need to know how much information people actually have, or if there are information asymmetries, and so on. In that sense, when doing ethical philosophy one needs to be rather more rigorous in one’s language than in the context of a scientific theory, no matter how much math may appear in the scientific theory.
Moving on, let’s imagine a different situation in which A notices B does not appear to be acting rationally. B may have all the information in the world, so A is satisfied on that account, but let’s say B’s thinking seems impaired or incorrect in some way that makes A wonder about B’s state of mind and concerned B may not be responsible in a normal or usual way for what B is choosing, maybe B is addicted to some substance in apples, maybe B has incorrect beliefs relating to what happens when he or she eats the apple (which could, of course, also be an information problem, but could be an illogical or irrational generalization as well), maybe B read something (presumably by someone no longer alive in our little one person world) that applied psychological manipulation that predisposed B to eat apples but on a basis A suspects B would not care about if the psychological manipulation were explained or addressed with some counseling or countervailing arguments or what have you. Doesn’t really matter for our purposes, the main point is A believes he or she has some reason to doubt B is thinking rationally when expressing his or her preferences.
Here again we’re in an analogous situation to full information. If A doesn’t really care whether B is acting rationally or not and, in particular, does not feel he or she has any ethical responsibility to help B by rendering his or her preferences consistent with what A imagines a rational person might do, then A can would go along with the proposition he or she should maximize total social utility. If A does care, then he or she may reject the proposition he or she should maximize total social utility. Again, A could preserve his or her support for the proposition in question by adding some sort of stipulation or condition that addresses this complication, for example, some condition that we should maximize total social utility as long as the people involved, B in this case, appear to be acting rationally. As with information, this case is interesting because an assumption of rationality is often considered part of economic theory, not as a condition of a particular market structure as with information, but as a condition for economic models and theory in general. However, again, it is typically addressed in a scientific context as a simplifying assumption that can be left either true or false but, again, in terms of evaluating the ethical proposition we should maximize total social utility the issue of the degree of rationality of those involved appears as a factual premise the truth of which is crucial to the our evaluation of that proposition. In a philosophical and ethical context we need to know, exactly, what we’re saying about rationality. Are we saying everyone is always perfectly rational or rational to some particular degree, some people are perfectly rational or rational to some particular degree, everyone is sometimes perfectly rational or rational to some particular degree, etc. If we’re going to get serious about evaluating the ethics involved, we have to get rigorous in terms of defining the argument we’re actually making.
Let’s do one more example of issues in this one person world (plus an observer) case, say a few quick words about the multi person case, and then call it a day, shall we? Another complicated situation in this one person no interpersonal conflict case is what happens when we allow the possibility of future generations. We still arguably have only the one subject to think about, but we have more on the way. Let’s say B is thinking about eating the last apple and A has every reason to suppose future generations would have also enjoyed eating applies if Little Piggy, sorry, B didn’t eat the last one. One issue is whether this situation counts as a sort of interpersonal conflict that would render the proposition we should maximize total social utility irrelevant because we cannot make interpersonal utility comparisons. Could A still support the proposition we should maximize total social utility in this case without a second thought because it has no implications in this situation because of the interpersonal conflict and A would thus be free to intervene of behalf of future generations or not on the basis of some other ethical argument? It’s not entirely clear. However, I think the conventional view is probably that utility in economic theory is only defined with respect to the preferences of actual living people who can express them via their choices, or the internal perceptions of people who are actually alive to experience them if you’re using that other definition of utility, so future generations would really have no standing in terms of utility, and this situation would not represent an interpersonal utility comparison, per se. As such, A would again have to think about whether he or she agrees with the proposition he or she should ignore his or her assessment of the likely preferences of future generations and whether or not they would be harmed by B expressing his or her own preferences. Again, A may want to reject the proposition we should maximize total social utility on this basis. Alternatively, A could try to maintain support for the proposition by adding some stipulation or condition, such as we should maximize total social utility after accounting for the interests of future generations.
What happens when we start adding more people to our little world? A is now considering not just B but the rest of the alphabet as well. In that case we not only open the door to interpersonal conflict but we lose the capacity to know conclusively when interpersonal conflicts are present or not. The more people we add the more likely someone may have ethical beliefs or concerns about what B is doing that would generate an interpersonal conflict and render the proposition we should maximize total social utility irrelevant. A could support the proposition we should maximize total social utility with no problems, but only because the proposition would be rendered empty and devoid of practical implications, thus allowing A to resolve the interpersonal conflicts on whatever basis he or she found reasonable. Alternatively, A could try to maintain some practical implications to the proposition we should maximize total social utility by adding some condition relating to the standing of other people, maybe something like we’re not going to give other people standing as far as registering an interpersonal conflict unless the conflict involved reaches some arbitrary level of directness or intensity or concreteness or what have you. Otherwise, I’m afraid we pretty much have to say the proposition we should maximize total social utility has no practical implication in realistic situations.
So what have we learned from this laborious analysis of the implications of the proposition we should maximize total social utility when that proposition is not combined with other unrelated ethical propositions relating to property rights and market mechanisms designed to resolve interpersonal conflicts on the basis of economic power in markets? Well, one thing we discovered is that it’s not as ethically uncontroversial or philosophically innocuous as economists seem so eager to suggest. There are a few rough bits. One issue is that those accustomed to utilitarian ethical theories in which utility plays such a major role will naturally look to the concept of utility to find the ethical content of neoclassical welfare economics, but when they look at utility as used in economic theory they will find only a small portion of the ethical or normative content of neoclassical welfare economics. Another issue is that the proposition we should maximize total social utility can easily become ethically controversial in practical situations unless we start adding stipulations or conditions to head off the potentially controversial bits, and although those stipulations and conditions sometimes appear in economic theory in one guise or another and one context or another, they don’t appear in the form and context in which they really should as far as the ethical arguments go. I think it’s fair to call this initial, fundamental level of ethical thinking the core of the onion of bad economics because the defining feature of bad economics for me is the presence of opaque, unclear, unsystematic ethics, and that seems certainly what we have here.
By the way, if one likes, one can do as I did in my little book and suppose a final ethical implication of accepting the definition of utility used in neoclassical economic theory and the ethical proposition we should maximize total social utility is that economists qua economists will not take up the ethics of resolving interpersonal conflicts; however, that one doesn’t seem entirely correct given the addition of other ethical propositions, not involving utility, introduced into economic theory later, some of which may be part of neoclassical welfare economics proper and some not. The relevant ethical proposition is probably something more like economists qua economists will not take up issues relating to the ethics of resolving interpersonal conflicts on the basis of utility, although they may address them or partially address them on other bases. (I say partially because of distributional indifference.) Not sure this adds anything to the mix, but explicitly introducing such a proposition would help one keep in mind that the word “utility” in neoclassical economic theory is there mostly to serve as a red herring and preclude the broader use of utility in an ethical context and in particular in the context of resolving interpersonal conflicts, thus demonstrating that neoclassical welfare economics is not a utilitarian ethical theory as much as an anti-utilitiarian partial ethical theory, given the importance of resolving interpersonal conflicts in any ethical system.
