Last week I tried to push myself a bit by talking about a critique of neoclassical welfare economics I’ve always rather struggled to understand. As fate would have it, a helpful commenter on the Twitter Wednesday afternoon tweet storm version of that blog post promptly shared a lecture dealing with an entirely different perspective on neoclassical welfare economics I’ve always rather struggled to understand. How does the old adage have it? It never rains but it pours? But I’m just joking around. I actually enjoy trying to work through things I find initially puzzling. It’s a peculiarity of mine. Like doing a little logic problem. So let’s take up this issue of the popular, perhaps even orthodox, positive take on normative neoclassical welfare economics, the world of illustrious residents of the Villa of Academic Economics like Samuelson, Arrow, and Sen.
The link in question was to The Econometrics Society (Yale) 2020 Frisch Memorial Lecture by David Pearce (NYU). It’s about an hour long, so there’s a modest time investment involved. I’m generally a proponent of the written word. Talking heads make me exhausted. I’m always trying to hit pause and go back. Wait, who said what now? But I found it worthwhile viewing. Anyway, I hope my commentary will be sensible even without a prior viewing, if one isn’t quite up to that just now. I put a reference at the end of this post. One of the nice things about the lecture is that it gives a little historical overview of that stream of academic economics that purports to be a positive study of normative neoclassical welfare economics and related “choice theory,” how it all started, where it is now, and an indication of where some economists anyway would like it to go in the future. At its most general level, the theme of the lecture for me is the stultifying, oppressive weight of an existing literature when combined with an academic determination to incrementally advance that literature rather than going back and reconsidering fundamentals. But maybe that’s just me. I’m sure there’s a lot more going on in the Villa of Academic Economics than just that particular dynamic. Normally, of course, I avoid the Villa as much as possible. Too precious and clever by half for a simple man like me. Too hidebound and myopic, as well. But I know some people like that sort of thing, so I suppose I can visit, if only for the afternoon.
As I noted above, I’ve struggled in the past to understand the ostensibly positive, scientific analysis of what seems plainly the evaluative, normative theory of neoclassical welfare economics, which clearly purports to evaluate economic systems and outcomes, explain what is optimal, tell people what they should do as far as economic policy. I could never really make out exactly what these particular economists thought they were really doing. But I think now I may be getting a glimmer. I think what’s going on is economists working in this tradition don’t think they’re evaluating anything, recommending what anyone should do, or indeed putting forth any substantive normative propositions at all, or at least they aren’t willing to publicly acknowledge doing any of those things, anyway. They see themselves as simply analyzing certain given, essentially arbitrarily defined issues and concepts under a certain given, essentially arbitrary set of conditions in a purely positive way. They’re about what’s “optimal” given these arbitrary definitions, conditions, and constraints, not what’s optimal in some generalized, commonplace, ethical sense. In their minds, they’re all about normative propositions relevant to the essentially arbitrary intellectual construct or artifact I call the Fairy Land of Economic Theory, not at all about normative propositions relevant to the real world. If other people choose to attach normative significance in the real world to the normative in form only conclusions of neoclassical welfare economics meant for the Fairy Land, choose to interpret “optimal” as optimal, maximizing “utility” as something desirable, that’s on them. It has nothing to do with economists or economics.
On that bit of confusions rests, I suspect, a great deal of what I call bad economics. Obviously, an empirical science about empirical preferences and a logical analysis of arbitrarily defined terms under arbitrarily defined conditions could never tell us what is ethically optimal in the real world. If we’re talking about normative propositions about individual preference rankings under a condition of no interpersonal conflict, well, someone may have an observable, empirically verifiable preference to unwittingly eat a poisoned apple in the real world, but what, if anything, we should do about it is a normative or ethical issue, not a positive issue. If we’re talking about internal perceptions of satisfaction from preference fulfillment in the real world, I suppose we might look at brain scans or whatever to define or get data on intensity of preference or perceptions of satisfaction. However, again, if we’re proposing attaching normative significance to intensity of preferences or intensity of perceptions of satisfaction from preference fulfillment in the real world, however defined or measured, that’s a normative or ethical issue, not a positive issue. Someone may have a verifiably intense preference to hit his or her depressed, nearly suicidal neighbor on the head with a rock, but what, if anything, we should do about it is a normative or ethical issue, not a positive issue. A megalomaniac may have a verifiably intense preference to enslave the world that let’s suppose is verifiably greater than the world’s combined preference to not be so enslaved, but what, if anything, we should do about it is a normative or ethical issue, not a positive issue. It’s ethics. It’s not science.
I tried to think of an example in a context other than economics to illustrate what I think these economists think they’re doing, and I came up with the following scenario, which I think works well enough, although I haven’t spent a huge amount of time thinking about it and something might always occur to me later. Usually does. Anyway, imagine some people decided to delve into the collected random scribblings of their dear departed neighbor, Joe Blow, in an effort to discern what Mr. Blow considered ethically “optimal” in those scribblings. They’re working in the context of a normative or ethical theory, but they’re not doing ethical philosophy, per se, not trying to evaluate Mr. Blow’s ideas, endeavor into the normative or ethical inputs, question the factual premises, assess the degree of ethical controversy, normatively evaluate the conclusions. No, they’re simply looking at positive, empirical data, the scribblings of Mr. Blow, and applying logical analysis to those scribblings, to try to answer the positive, empirical question: What did Joe Blow consider ethically optimal, according to his scribblings?
A couple of issues or questions leap immediately to mind. First, are they doing “science?” There’s a theory of sorts, a theory of what Mr. Blow meant to say is ethically optimal in his scribblings. And there are empirical data involved. But assuming the entire archive has been analyzed, there’s no real empirical prediction or testing of the sort one associates with conventional notions of empirical science. Second, why are they doing it? Absent any implication there’s something compelling or interesting about Mr. Blow’s ideas in this area, some merit to Mr. Blow’s ethical ideas, why are they taking up that particular research question? Why are we meant to care what Mr. Blow’s scribblings say? What’s it all about? Who gave them this positive research project? A normative theory without normative analysis and normative evaluation is basically worthless. Simply knowing what Joe Blow has to say about ethics said has no normative significance, per se. If we were wondering whether there was any value in Mr. Blow’s ideas as ethical philosophy, we’d have a whole lot more work to do, indeed in many ways the lion’s share of the work for that particular project. These are the sort of thoughts that occurred to me while listening to the lecture on the positive approach to normative neoclassical welfare economics, and some other thoughts followed close upon those. Let’s discuss them a bit in the context of the lecture itself.
The tale told in the lecture begins in the 1920s with Mr. Frisch’s efforts to turn economics into a science “like physics” (of course), while cryptically warning against trying to “unreflectively” apply the methods of a science, like physics, to economics. This inauspicious start soon bore funny fruit as our aspiring scientist friends collided with those great proponents of science, the logical positivists, and their notion that normative or ethical statements are “nonsense,” that is, not meaningful, in a scientific context because they are not falsifiable. Of course, the insight of the logical positivists is correct in the context of science. Ethical propositions are not “falsifiable.” One can’t do ethical philosophy using the scientific method. One cannot derive an “ought” from an “is.” Normative is different from positive. There is no objectively, empirically correct ethics. No, what caused problems with logical positivism was not that proposition, per se, but misstating the status of “science.” Humans have subjective moral sentiments. They have a drive, a need, a desire to be ethical, to treat other people ethically, to be treated ethically, to live under a social system that exhibits good ethics, to do good and oppose evil. The error of the logical positivists was not their analysis of normative propositions in the context of empirical science, but the proposition the only thing intellectually important or relevant to humans and human society is empirical science.
Unfortunately for the field of economics, economists were unprepared to give up talking about normative issues, evaluating economic systems, dispensing advice, or seeming to anyway, even though it was “nonsense” in the true science they admired so much. Basically, they wanted to have their cake and eat it too. This is what may have led some economists to try to transform neoclassical welfare economics from an ostensibly substantive normative endeavor to a Joe Blow research project about arbitrarily defined concepts, conditions, constraints, in other words, to substitute an arbitrary, random, academic “optimal” for a real optimal. Economists in this tradition were no longer interested in actually evaluating the normative inputs or outputs of their normative theory, checking their relevance for the world, the degree of ethical controversy involved. To paraphrase Dr. McCoy, they were scientists, dammit, not ethical philosophers.
Soon enough they ran into the problem that absent normative evaluation of their normative theory their positive research project was underspecified. No one was entirely sure what they were meant to be analyzing or why. They had transformed a straightforward normative discussion into a vague, opaque positive one. In such an odd situation, it was only natural for critics to suppose the only criteria economists were using to determine what normative concepts to positively study were economists’ own uninformed, unexamined, unevaluated, unreported normative beliefs in their role as non-economists. To return to our example, imagine our researchers uncovered a trove of artwork by Mr. Blow. Was the project meant to be an analysis of his scribblings alone, or all his output, including his artwork? What about the material from Mrs. Blow about Mr. Blow and his ideas? Was that meant to be part of the project as well? What exactly was the positive research question? Unfortunately, no one wrote it down. There was no reference, no work order, no mission statement. Some people wanted to address one question, some another. “Why are you analyzing that? I don't know, aren’t we meant to? Why not analyze that? Who cares?”
Economists like Samuelson and Arrow added to the confusion by looking at the older inaccessible perception of preference satisfaction definition of “utility” and opining normative content going beyond the observable facts of revealed choice were “meaningless” in the context of their positive analysis of normative concepts. It’s fine. Why not? That’s one way to define “utility.” I explain often how different definitions of “utility” imply different normative propositions serving as inputs to neoclassical welfare economics, are more or less ethically controversial, but recall ethics was all meant to be nonsense to our wannabe scientists. They weren’t thinking of that at all. Ostensibly, anyway. However, there are many ways to think about “utility,” preferences, welfare, satisfaction, and all the other terms and concepts involved, and some economists wondered why we weren’t discussing the positive, empirical data relevant to some or all of those instead. In a normative analysis, it’s quite obvious why someone might support an ethical half-theory eschewing resolving interpersonal utility comparisons or not, one type of “utility” or another, normative evaluate “utility” under one set of conditions or another. But from this positive perspective, there’s no real reason to study anything in particular.
To close out our discussion for today, here are some other notable bits from the lecture. When Arrow discussed the limitations of an ethical half-theory based only on observable individual preference rankings, under the guise of a positive analysis of the type of “utility” he accepted as his mission to analyze, it apparently led some to perceive the “death of neoclassical welfare economics.” That was circa 1978. How one kills a positive analysis of concepts is anyone’s guess, but one can perhaps find a clue in Sen’s statement that Arrow’s Theorem “generated further pessimism in an already gloomy assessment of the possibility of a reasoned and satisfactory welfare economics.” Not sure "reasoned and satisfactory” are common evaluative criteria of positive science, but there you have it. Seems likely there’s some funny business going on not only in determining the research project, the random normative concepts and issues to be positively analyzed, but in determining what that research project is meant to accomplish, how to evaluate it. In due course, our scientist friend, Dr. Arrow, is reported to have said, “most ethical systems purport to divine who is the deserving one.” It’s an odd word choice suggesting some lingering confusion about the distinction between normative and positive. No one divines anything via ethical systems. Ethics is not about making empirical predictions. Ethics is about people talking about their moral sentiments, trying to formulate and explain rational systems based on those moral sentiments, making ethical decisions and trying to communicate and reach common ground with others. It’s not science; it’s philosophy. Our guide reports Mr. Arrow’s comment as being remarkable, noting it’s not as though economists were “trying to divine the appropriate choice of the next Dalai Lama,” they’re “simply trying to allocate some resources.” Note “divine,” again. And simply trying to allocate resources? Doesn’t seem a very simple project to me, if I’m being honest, and certainly the stakes are potentially quite high, indeed, life or death for some. Our guide then makes a case for economists to do more work analyzing a cardinal form of “utility” incorporating intensity of preferences, but not a normative case, a positive case based on the notion we have empirical information, data, seemingly relating to preferences we’re not utilizing if look only at “utility” defined as ordinal individual preference rankings. The tale ends with modern choice theory and the search for voting systems conforming most closely to what our guide suggests is the “the ideal,” defined relative to what ostensibly arbitrary and meaningless subjective ethics I’m not entirely sure.
Hopefully, modern choice theory will be somewhat less conducive to bad economics and folk economics than traditional neoclassical welfare economics, but certainly that latter theory is still with us as well, and we know well the sort of problems that can cause. And so on it goes. I think I’m starting to have some insight into this tradition, but honestly, if I had to discuss normative neoclassical welfare economics the confused, inapt, “positive” way they discuss it in the Villa of Academic Economics, I suspect I’d drink a glass of sparkling wine and throw myself from the balcony straightaway. No, I’ll take my lonely mountain path. I’ll wage my battle against bad economics there, in the brisk air and unforgiving light of the bright afternoon sun, not in the cramped and oppressive atmosphere of the Villa.
References
Individual and Social Welfare: A Bayesian Perspective. Frisch Memorial Lecture. August 12, 2020. David Pearce. https://www.youtube.com/watch?v=upJadnNxxlQ.
