Digressions&Impressions

RIP: Noel Swerdlow (1941-2021)

My graduate school classmate at The University of Chicago, Matt Kerrigan Frank, alerted me to the fact that Noel Swerdlow recently passed away. (See this obit.)

I met Noel on one of the first days in graduate school back in September 1995. The seminar room was busy filling up with faculty and graduate students from philosophy and history of science for one of the many, weekly joint seminars I would attend during the next seven years. I was trying to impress David Malament with my knowledge of the Huygens/Newton debate which I had been researching under the guidance of George Smith (one of my undergrad gurus). An old crank -- I now realize he was then barely older than I am now-- interrupted my narrative to tell me that Huygens' objections to Newton's inverse law were primarily metaphysical. I tried to explain to the guy that Koyré's view had been refuted (by George and me). I never get a chance. I asked David, "who is that old man?" and he informed me it was Noel Swerdlow.

Over the next few years I saw Noel in action at the various joint seminars: often abrasive and impatient, but at times quite brilliant. Some of his exchanges with visiting scholars could be quite spirited. (His exchange with Motti Feingold was legendary, but they were or became close friends.)

My relationship with Noel evolved once he spent some time away at Dibner (where George had become director). After we met again on campus, he told me with a big smile "you could have told me about all the archival research that went into the Huygens paper [see here]. I thought you were just bluffing. It's really marvelous work--I love the argument with the maps." I meekly tried to say, "i tried." But something had changed, I was now on his good side. And to receive a compliment from Noel (or anyone in graduate school) was one of the highlights of my graduate career.

He was famous for his work on Babylonian astronomy (he was a student of the legendary Otto Neugebauer), but I have to admit that as a PhD student I loved learning about Copernicus and his (obscure to me) peers from Noel. I also have a memory of him playing piano, but I have not seen this mentioned in any obituary so maybe I am confused. We discovered our shared love for opera at the Lyric, and I always looked forward to seeing him there. (He was very learned on opera.)

When I was nearly done with my PhD, I had the luminous idea to ask him to read Adam Smith's posthumous "History of Astronomy" with me. This piece was central to my dissertation. I was delighted he agreed, and even more that he had never read it before. (I always assumed Noel pretty much had read everything.) I expected that Noel would be extremely critical of Smith as a historian (as the editors of the Oxford edition are). But Noel was, in fact, quite impressed with Smith's historical acumen and, more importantly, gave me a graduate seminar on all the possible sources Smith could have consulted during the 18th century. He also flattered my vanity in assuming that I was intimately familiar with all the works of Regiomantus, Riccioli, and Peuerbach. Often after our reading group at The Medici I had to go to special collections to figure out what he was saying. (This was before the age of google.books!)

Later, when I realized I had bombed out on the job-market first time around, he even helped me apply (and wrote a letter) for a project on Smith as historian of astronomy. But that was not funded.

At the time, i assumed it was normal for historians of astronomy to have a position in an astronomy department.  So, I later used that as an argument for my paper on "philosophy and a scientific future of the history of economics" But I have learned it is much less common (although archeoastronomy is sometimes practices by professional astronomers). 

That paper was part of a rather contentious symposium. organized by Paola Tubaro & Erik Angner. To be clear the symposium was lovely but our audience of established historians were underwhelmed. I think referees managed to prevent bublication of some of the papers.
Anyway, my core idea was that the history of economy theories could play an evidential role in present day economics as the custodians of the theories and data of the past. This was not well received at all. But it let to an invitation by Bertram Schefold to reflect on the possibilities and significance of economic history. I did not know much about the practice of economic history. But as I was musing about the invite, I wondered what the oldest economic data might be. The hunch being that if you have data very far apart this might help you to calibrate important measures.  So on a lark I googled (!) "Babylonian economics" and I was directed to Alice Louise Slotsky The Bourse of Babylon Market Quotations in the Astronomical Diaries of Babylonia, which I read and loved.

I was astonished to learn that the very clay tablets with astronomical data Noel had studied also had economic data on them! I started drafting one of the craziest and fun papers I ever wrote, "Prophecy, eclipses and whole-sale markets: A case study on why data driven economic history requires history of economics, a philosopher’s reflection." [published in Jahrbuch für Wirtschaftsgeschichte/Economic History Yearbook 50.1 (2009): 195-206.]

With some trepidation I sent an early draft to Noel. Rather than scolding me for my amateurishness (this was the initial response of some of my Leiden colleagues who were experts on ancient Mesopatamia--I think they really thought I was crazy), he wrote lengthy letters on the quality of the data that I wanted to use. (The published version of the paper includes excerpts from his letters to me. ) Nobody seems to have read this paper, but it was good enough for me that Noel (who called my attention to relevant other scholars) took it seriously enough to try to help me improve it.

In later years, Noel would always read works I sent him. Sometimes even my blog. Even when I relied on impeccable scholarship of others, he would correct mistakes that had crept into now famous scholarly works. My favorite letter, however, is this one-liner, "Take it down; it's nonsense!--Noel."

On the Convergence of Economics and Computer Science (with comments on the role of moral reasoning in AI)

We can generalize this idea across our entire model space. Suppose we enumerated all the possible numerical pairs <error, unfairness> achieved by the models we are considering (e.g., SAT cutoffs)....

So each point corresponds to a different model; the x-coordinate of the point is the model’s error, and the y-coordinate is its unfairness score...The key thing to realize is that any model that is not on this boundary is a “bad” model that we should eliminate from consideration, because we can always improve on either its fairness score or its accuracy (or both) without hurting the other measure by moving to a point on this boundary.

The technical name for this boundary is the Pareto frontier or Pareto curve, and it constitutes the set of “reasonable” choices for the trade-off between accuracy and fairness. Pareto frontiers, which are named after the 19th-century Italian economist Vilfredo Pareto, are actually more general than just accuracy-fairness trade-offs, and can be used to quantify the “good” solutions to any optimization problem in which there are multiple competing criteria. One of the most common examples is the “efficient frontier” in portfolio management, which quantifies the trade-off between returns and risk (or volatility) in stock investing.

The Pareto frontier of accuracy and fairness is necessarily silent about which point we should choose along the frontier, because that is a matter of judgment about the relative importance of accuracy and fairness. The Pareto frontier makes our problem as quantitative as possible, but no more so.

The good news is that generally speaking, whenever we have practical algorithms for “standard,” accuracy-only machine learning for a class of models, we also have practical algorithms for tracing out this Pareto frontier. These algorithms will be a little bit more complicated—after all, they must identify a collection of models rather than just a single one—but not by much... 

While the idea of considering cold, quantitative trade-offs between accuracy and fairness might make you uncomfortable, the point is that there is simply no escaping the Pareto frontier. Machine learning engineers and policymakers alike can be ignorant of it or refuse to look at it. But once we pick a decision-making model (which might in fact be a human decision-maker), there are only two possibilities. Either that model is not on the Pareto frontier, in which case it’s a “bad” model (since it could be improved in at least one measure without harm in the other), or it is on the frontier, in which case it implicitly commits to a numerical weighting of the relative importance of error and unfairness. Thinking about fairness in less quantitative ways does nothing to change these realities—it only obscures them.

Making the trade-off between accuracy and fairness quantitative does not remove the importance of human judgment, policy, and ethics—it simply focuses them where they are most crucial and useful, which is in deciding exactly which model on the Pareto frontier is best (in addition to choosing the notion of fairness in the first place, and which group or groups merit protection under it, both of which we discuss shortly). Such decisions should be informed by many factors that cannot be made quantitative, including what the societal goal of protecting a particular group is and what is at stake. Most of us would agree that while both racial bias in the ads users are shown online and racial bias in lending decisions are undesirable, the potential harms to individuals in the latter far exceed those in the former. So in choosing a point on the Pareto frontier for a lending algorithm, we might prefer to err strongly on the side of fairness—for example, insisting that the false rejection rate across different racial groups be very nearly equal, even at the cost of reducing bank profits. We’ll make more mistakes this way—both false rejections of creditworthy applicants and loans granted to parties who will default—but those mistakes will not be disproportionately concentrated in any one racial group. This is the bargain we must accept for strong fairness guarantees. Michael Kearns & Aaron Roth (2019) The Ethical Algorithm: The Science of socially Aware Algorithmic Design, Oxford University Press, 80-84 [the linked page should have their diagram]

A few weeks ago, after I started blogging (here and here) about the ethics of algorithms,* I received a Facebook advertisement-notification from Amazon that Michael Kearns and Aaron Roth (the authors of The Ethical Algorithm) had become Amazon scholars. "What an interesting way," I thought, "to advertise books to me." I clicked on the link and read the corporately sponsored interview with them. Perhaps because my expectations were low , but I was surprised how thoughtful -- "the main thesis of our book, which is that in any particular problem you have to start by thinking carefully about what you want in terms of fairness or privacy or some other social desideratum, and then how you relatively value things like that compared to other things you might care about, such as accuracy" -- they sounded and how relevant to my own project. It was clear theirs was a project about thinking about trade-offs (a favored phrase of theirs) about social values in algorithm design. More important to me, since I had just argued that thinking about the ethics of algorithms seems to replicate the very social problems familiar from thinking about ethics in economics, I wanted to see if my hunch was correct. Two clicks later I ordered the book, and a day later it was delivered.

A few years ago (well back in 2012), I used to joke that as even theoretical economists become more applied, becoming (thanks to increasingly cheap computer power) experts at massaging results out of giant "administrative data-sets,"+ economics ran the risk of being displaced by statisticians and, especially, computer scientists. While then I was largely ignorant of deep learning, it was clear to me that fruitful searches for robust and surprising correlations in data didn't require (as Justin Wolfers then suggested) the restrictive assumptions of economic theory. (Think about it: econometrics and economic theory artificially restrict the search space that can be more fully explored with deep learning machinery.) And indeed deep learning can dispense with economic theory.**

But in reading their interview and their splendid The Ethical Algorithm I realized that back in 2012, I had missed a crucial issue: that algorithmic design is conceived in terms of optimization problems under constraints. Since Lionel Robbins (back in the 1930s) this just is the definition of economics (and it enabled a split between ethics and economics). And once in algorithmic design you are interested in more than predictive accuracy, and so have to deal "with multiple competing criteria," a whole set of mathematically precise diagnostic tools familiar from economics can be imported into algorithmic design (as the passage quoted above suggests). So, I now realize that increasingly computer science and economics will merge (as presumably is happening already in finance-notice the passing mention of "portfolio management" in the block quote above). 

The previous is sufficient for a digression. But I also want to call attention to how moral issues are conceived, and not, in the (ahh) paradigm articulated by Kearns and Roth. And, in particular, I want to call attention to four levels, or sites, where they enter in. So, formally moral issues enter into two sites of algorithmic design: (i) first a choice in setting a goal (such as privacy or fairness) or success criterion for the algorithm. In general, Kearns and Roth view this as a constraint on accuracy that generates a trade-off among accuracy and other possible criteria. Kearns & Roth are very good at explaining that the way one makes precise these moral goals need not be univocal or can be very context sensitive. (So, there are different ways to think about fairness or privacy in algorithmic settings.) 

In turn, these trade-offs can be modeled in terms of pareto frontiers. And this generates the second site: (ii) at the level of a decision about ''which model on the Pareto frontier" to use in practice. Conceptually that's a distinct choice from how to encode 'fairness' into an algorithm, but obviously one can imagine that in practice, in the spirit of experimentation, there are going to be interactions between (i) and (ii) within a general goal-oriented design process. (Recall that an algorithm can be understood, even identified, in terms of the functions and goals it serves--something I have promised to return to.) 

Some time soon, I hope to write some more critical posts about their hope of turning algorithm design into a kind of dream of turning ethics into an optimization problem. (That way of putting it is indebted to Kathleen Creel, a young scholar who has written a gem of a paper on opacity in computational systems.) And one can see how the combination (i) and (ii) lend themselves, despite the situationism in (i), to a form of moral reasoning familiar from (and analogous, if not identical to) utilitarianism. Of course, one can let other values enter into the decision of (ii). 

Now, in their argument (ii) tends to happen in algorithm design/development process within, say, a company. But the choice one makes about which model on the Pareto frontier to inhabit generates (ahh) consequences to wider society. Some of these wider consequences -- laws, regulations, social norms, etc. -- are already internalized as constraints, but some are simply outside the company's mission/attention and not anticipated by the legislature.

The point in the previous paragraph fits their general argument. Because the "core concern" of their book is that "optimizing" on some explicit social goals, predictably and foreseeable, has (even in the relatively short term) unintended and perhaps ex ante unknowable side effects. (188) There is, thus, not just (as Kearns & Roth note) a (iii) a prior moral and political decision to be made whether or the extent to which algorithms are permitted or primarily responsible for decisions in a domain or (as they often describe it) norm enforcement (177). In the book they offer as an example, of the former "automated warfare," (175; 178).  But there is also (IV) a question how society should think about, and perhaps compensate for or mitigate, these entirely unintended and perhaps unknowable side effects that are a foreseeable outcome pattern from embracing algorithmic practices. This matters especially if such outcome patterns (recall) create  asymmetric harm patterns to vulnerable populations.

And, not unlike the economists, the way Kearns & Roth have set up their conceptual scheme (IV) turns out to be delegated to policy and so is not part of the ethical algorithm at all (or the responsibility of the firms that profit from them). To put this in their terms, we (now) know that a system with the very best portfolio management can create unexpected, general externalities. Something similar is now foreseeable in their field. (I am not sure what the equivalent term of true Knightian uncertainty in AI is, but they need it!) So given (IV), it would be good to prevent regressive forms of socialization of risk while privatizing profit.

PPE as Synthetic Philosophy (or synthetic science)

PPE (Philosophy, Politics, and Economics) is an increasingly popular undergraduate and graduate major. In practice there are two models: one is what I like to call a confederate model: in this approach students take a mixture of courses from the three contributing disciplines. One is what I like to call a federated model: in this approach students take a mixture of courses from the three contributing disciplines and then jointly (or partially) bring these to bear on a capstone topic (a paper, thesis, project, etc.) Most PPE programs I am familiar with fall somewhere between these two models.

Obviously other options are possible, including the reflexive option in which students study the three contributing topics, in part, from each others's perspectives (i.e, philosophy and politics of economics, the economics of philosophy, etc.) A final approach would be the integrated model in which students take some courses from the three contributing disciplines, but primarily are learning to apply them jointly  (as mutually supporting, mixed methods). Obviously these approaches can be blended in various ways and the list is not exhaustive.

Recently I visited the third annual PPE conference in New Orleans, and to my surprise I came away thinking that PPE may be a viable research area. I think this for a number of sociological reasons: first, the number of PPE programs is growing and so are their enrollments.* This means that there is steady hiring in PPE. Second, while the (quite popular) confederate model means one can hire faculty without much interest in PPE to teach within PPE, hiring folk with fondness and interest in PPE prevents downstream troubles. These two reasons are, we may say, demand-led.

But third, what struck me hard at the PPE conference is that there is now a critical mass of people ((ahh) the supply side) that are confident in offering and evaluating arguments and proposals with competence derived from at least two of the three contributing areas, while maintaining (a) shared conversation, and (b) being capable of informed and overlapping criticism(s), and (c) without disagreements wholly being methodological. In addition, while the conference attracts quite a number of market friendly types --  this became clear during a straw-poll taken by Al Roth during his lecture on kidney selling --  PPE is clearly not synonymous with 'libertarian friendly' or 'heterodox economist.' Quite a few sessions explored criticism of really existing capitalism.+

As an aside, during the past year I read (recall) James Scott’s Seeing Like a State, which gave me a (relatively) contemporary exemplar in what PPE research could be. This autobiographical example is connected to what take to be one of the main problems with the (fast growing and developing) PPE society: political science is primarily represented by political theorists (and philosophers masquerading as political theorists) and that there is little contribution from the rather large group of political economists in political science departments (of a sort that publish in journals like Review of International Political Economy, New Political Economy, and European Journal of Political Economy among others). That philosophers and economics find it easy to talk to each other is no surprise -- we share overlapping methods and research topics (decision theory, Bayes, justice, welfare, future generations, etc.), even overlapping research cultures.

But (inter alia) political scientists bring facility with mixed methods to the table as well as a focus on power. Political scientists are also naturally at ease with theorizing about far-from-ideal and especially messy circumstances (in ways that P&E are not).** Non-trivially, political scientists also have much better developed research ethics toward would-be-research subjects (while philosophers and economists tend to assume that research ethics is limited to not getting caught plagiarizing [okay that (see here) may be changing]). 

So, what could PPE research be? A lot of it will be folks doing research in one of the three core disciplines while drawing on or engaging with one of the others (much of what I do fits in this category). But here's one possible, more ambitious answer: something akin to synthetic philosophy. Recall that:

Synthetic philosophy [is] style of philosophy that brings together insights, knowledge, and arguments from the special sciences with the aim to offer a coherent account of complex systems and connect these to a wider culture or other philosophical projects (or both). Synthetic philosophy may, in turn, generate new research in the special sciences, a new science connected to the framework adopted in the synthetic philosophy, or new projects in philosophy. So, one useful way to conceive of synthetic philosophy is to discern in it the construction of a scientific image that may influence the development of the special sciences, philosophy, public policy, or the manifest image.

So, understood quite a bit of PPE is already (developing into) synthetic philosophy: work on climate justice, obligations to future generations, the institutions and norms of transition economies, institutions and norms of uncertainty, the political economies of migration are all fertile areas that mix normative theorizing and various social sciences.*** More subtly topics that have been sidelined in the constitutive disciplines may be quite important in developing PPE into synthetic philosophy. I give two examples, first philosophers are wont to think that poverty is exhausted by the difference principle; but in fact, poverty impacts all kinds of significant projects (migration, international law, hunger, intersection with gender and parenting, etc.). Second, philosophy of education has almost no status anymore within contemporary philosophy (sorry); but the political economy and philosophy of education matters a lot to development, migration, and future oriented projects within PPE.++ 

One may suspect that what makes PPE a viable research area is the felt need to combine normative theorizing (the strength of philosophy) with social science research (the strengths of Ps&E). I don't think this is mistaken; but it's not the whole truth. In part it is not the whole truth because much of the theory required  -- that is normatively sensitive and sensible [and avoids noxious status quo bias] -- is still to be developed within such social science(s).+++ So, it's not the whole truth because philosophy can contribute more than pre-existing normative theory. And in part because the social sciences supply more than just empirical knowledge/theory; Ps&E also supply a large range of methods and familiarity with policy environments. 

The previous paragraphs are, of course, a promissory note.

For those that worry that even PPE as (integrated) synthetic philosophy is philosophy hegemonic (or merely cleverly piggybacking on the social status of economics), I am open to the idea that the future of PPE really is Synthetic Philosophy Science.

On "Colonial Claims:" President Wilson, Kant, and Mill.

V. A free, open-minded, and absolutely impartial adjustment of all colonial claims, based upon a strict observance of the principle that in determining all such questions of sovereignty the interests of the populations concerned must have equal weight with the equitable claims of the government whose title is to be determined.--President Wilson January 8, 1918 Fourteen Points.

Much of President Wilson's Fourteen Points reads as if it is inspired by Kant's Perpetual Peace. (Some other time I'll say more about this and their shared program of peace based on international law, confederation, and commercial flourishing.) For the present purposes it is important to note it shares with Kant's blue-print a preference for self-determination, and a marked distaste for colonialism and imperialism. (That Kant was not always so Enlightened is by now a familiar fact.) For, regardless of Wilson's (well known) racism and hostility toward African American emancipation, Wilson's international program is explicitly aimed at "all the peoples of the world" and it embraces a program of "equality among the peoples of the world." (The latter quote is in the context of reassuring Germany that after its defeat it will not be subordinated.)  

Wilson's fourteen points were received as an anti-colonial manifesto by many colonized peoples. But Wilson's formulation introduces considerable ambiguity. Here I want to focus on two elements: first, how does one decide the "interests of the populations?" In particular, does one believe that the formerly colonized can articulate and decide their own interests. For example, John Stuart Mill infamously thought (recall) paternalistically that many nations were not ready for self-government. Mill's argument was not fundamentally racial, but rather based on cultural and economic factors in line with his broader theory of civilizational development.*

Wilson's fourteen points offers no hint of such a theory of cultural/economic development as a pre-condition to self-determination. All he seems to require is that (i) "every nation" can participate  on a basis of equality," as long as their "purposes are consistent with justice and the peace of the world." By 'justice,' I think he means, justice in its dealings with other peoples, that is, accepting international treaty obligations and natural right. This is clearer, perhaps in (ii) his claim that world "be made safe for every peace-loving nation which, like our own, wishes to live its own life, determine its own institutions, be assured of justice and fair dealing by the other peoples of the world as against force and selfish aggression." This (ii) is very important because unlike, say, Kant, Wilson does not require that the participating states are republics (in  Kant's sense, that is, what we would call 'liberal' or 'representative' democracies with division of powers). Wilson is clearly allowing countries to opt for non-republican institutions. (I take it this is a nice gesture toward then revolutionary Soviets, and various monarchic countries.)** It is notable that the quotes from Wilson talk of 'peoples' and 'nations' (not states); this is a clear rupture with the Westphalian (1648) framework.  

Second, and more problematic, what might he mean with the "equitable claims of the government whose title is to be determined." There are, in turn, two issues here. (A) Some diplomats at the time believed that this was a general limitation on Wilson's claim such that only the colonies of the defeated powers (Germany, Ottoman empire, etc.) would be involved. And these would then be subject of mandates/trusteeships. This is not entirely implausible given the later developments at Versailles and the League of Nations. But this does no justice to Wilson's wording. So, I think an  alternative is more likely.

Rather than seeing this as a limitation such that only the colonies of defeated powers are involved, it is more likely that (B) Wilson is thinking that the former colonial governments should be compensated for their would be loss.  After all, it was quite common then to think that reasonable expectations were the bedrock of justice, and that colonial masters had developed reasonable expectations about the returns on their property rights and substantial investment in colonized places. These expectations would now be violated and so deserved equitable compensation. 

The idea in the previous paragraph were familiar to Wilson's audience. For such compensation had been used to end slavery in the British West Indies in 1833. To simplify "The British government paid out £20m to compensate some 3,000 families that owned slaves for the loss of their "property" when slave-ownership was abolished in Britain's colonies in 1833."  Mill mentions it (without condemnation) in his Principles.**

Given Wilson's exact wording one may, in fact, see him as constraining the extent of such compensation. For, such compensation would have to be paid either by the newly liberated countries (a manifest injustice) or by the former colonial motherlands (as Britain did). But if the latter this entailed a massive transfer of income within the (formerly) colonial nations  (often from the working poor toward the colonial rich)--as it did in Britain. 

One may well object, and fairly object, that one shouldn't profit twice over from a manifest injustice. But the thought is an uncomfortable one politically because, as any American President surely realizes, the origin of property rights is (recall Spooner) rarely without blemish. 

 

 

 

 

 

On the Neuraths, Eugenics, and Translation

Now there is no denying that, given the trajectory of his later career, Neurath’s translation (with his first wife Anna Schapire-Neurath) of the second edition of Francis Galton’s Hereditary genius as Genie und Vererbung (1910) stands in need of explanation—indeed: contextualisation.¹⁷ Once that is provided, however, it becomes clear that his seeming ‘enthusiasm’ for eugenics was in fact considerably tempered. What prompted this tempering was, first of all, the complex influence of his father Wilhelm Neurath’s early arguments against Social Darwinism, albeit from an idealist perspective.¹⁸ In addition it must be noted that even the left of the day was by no means unanimously opposed to all eugenicist programmes and that the ideology of German eugenics before World War 1 was ‘less racialist than nationalist or meritocratic’, though unsavoury tendencies were already in evidence (Proctor, 1988, p. 20). Characteristically, Karl Kautsky dedicated the last chapter of a book of socialist reflections on Darwinism to the issue of ‘building a social consciousness in the spirit of racial hygiene’ (Kautsky, 1921, p. 265), explicitly in opposition to proponents of state-directed eugenic interventions like Wilhelm Schallmeyer. Finally, note must be taken of the protracted polemic against ‘the natural–scientific theory of society’ that Neurath’s mentor Ferdinand Tönnies conducted against Social Darwinists like Schallmeyer during the first decade of the past century in the journal of one of Neurath’s Berlin professors and dissertation examiners, Gustav Schmoller (Tönnies, 1925).

In light of this, the translation of Galton’s early work on eugenics may be regarded as an attempt to induce a certain sobriety into a highly topical public debate, especially since the translator’s ‘Preface’ draws the reader’s attention to what is significant about the second edition. Galton’s ‘Preface the second edition’—translated in full—not only stresses that the inheritance of acquired characteristics plays a still more minor part than originally thought, but also that in the meantime he had discovered the phenomenon of regression to the mean and the need to distinguish between genetic variations and so called sports. Galton concluded that ‘it is in consequence impossible that the natural qualities of a race may be permanently changed through the action of selection upon mere variations. The selection of the most serviceable variations cannot even produce any great degree of artificial and temporary improvement, because an equilibrium between deviation and regression will soon be reached, whereby the best of the offspring will cease to be better than their own sires and dams’ (Galton, 1892, p. xviii). That this radically reduces the scope for eugenics and reduces ‘the speed of the progress’ thereby promoted—as Galton put it in a passage from his Memories also reproduced in the translators’ ‘Introduction’ (Neurath & Schapire-Neurath, 1910a, p. vi)—is evident. It is surely not far-fetched to conclude that readers were meant to notice these consequences by the translators.

By 1913, in any case, Neurath observed: ‘One should clearly realise that instinct must fail with regard to the complex rational relationships created by the consciously shaped institutions of the social order and modern technology’ (Neurath, 1983, p. 5). So it is misleading to say that Neurath ‘in the years before the First World War became interested in eugenics’ (Proctor, 1991, p. 168) and wholly false to claim (as already noted) that it served as a ‘model and metaphor for social engineering’ (Richardson, 2009a, p. 19).¹⁹ Rather, by the time of World War 1, Neurath had lost interest in it because he realised that it was of no use for the transformation of society that he had begun to believe was required to alleviate the misery of the largest part of the population.--Thomas Uebel (2010) "What’s right about Carnap, Neurath and the Left Vienna Circle thesis: a refutation," Studies in History and Philosophy of Science Part A

Yesterday, while discussing LK Bright's work on the logical empiricist's taboo on (biological) racial explanation in human affairs, I alluded to an existing debate over the significance or meaning of the Neuraths's translation with his wife (a "professional translator") of Galton's Hereditery Genius. (As Galton makes clear in the preface to the second edition, by 'genius' he means 'natural ability,' although this does not rule out the impact of education and cultivation.) For even if one is friendly to the Vienna Circle it is a bit troubling to see one of its leading intellectual lights implicated in the active spread of eugenic ideas. (And one can imagine that at least some of those who are hostile will use it as critical ammunition.) Bright (and I) cite/link to a piece by one of the leading scholars of the Circle, Uebel (which I have quoted above). My own inclinations are with Richardson's position,* but what follows is agnostic about the relevance of early Neurath's views on eugenics to our interpretation of (a branch) of the later Vienna Circle.

First, I fully agree with Uebel that "even the left of the day was by no means unanimously opposed to all eugenicist programmes." In fact, as I remarked yesterday, many progressives, who want to improve society, after all, were invested in all kinds of scientific eugenics. (As regular readers know, I became aware of the significance of this by working on the overlapping history of philosophy and economics prompted by David Levy's scholarship.) Not all of these eugenic improvers were invested in racial hierarchy. But it is equally true that racial hierarchy is rarely fully absent from eugenic programs. As I remarked yesterday, Galton's book has a full chapter in which he presents his racialized hierarchy of peoples. The Neuraths's introduction does not challenge this classification. (It provides biographical background to his interest in the subject.) So, even if Uebel is right that the second edition of Galton's book does not promote eugenics, it does promote a racial hierarchy.

Second Uebel is right to call attention to the significance of the regression of the mean. The Neuraths's also call attention to Galton's comments (which he had added to the preface of the second edition) in their introduction. I want to make three comments about this. (a) The purported findings of racial hierarchy are not undermined by Galton's awareness of regression to the mean (which is a quality of a race); he repeats his views on the inferiority of "Negroes" in the introduction to the second edition (xxv-xxvi). In fact, (b) while Galton does not say this, one may even suspect that he thought this entrenches the hierarchy -- if one can maintain racial purity from the lower races -- because one can't breed up.

Now, it's true that Galton expresses (a lot more) pessimism about breeding for increasing existing characteristics in a population due to thinking through the implications of regressions to the mean (see, second edition p. xviii). But (c) he is explicit that it leaves open possibility for breeding for novelty or new characteristics (or what he calls "spots") which can then be passed on to the next generation. This can ultimately lead to a new "species" (race/kind). Galton goes on to call for the establishment of an experimental research lab in which breeding experiments on animals are performed alongside demographic study of human populations. So, a eugenic program that breeds new, desirable characteristics into a new kind of man may well be possible (and that's what, in fact, he calls for research into). Galton's remarks reveal the (evidential) research potential of drastic social engineering:

The great problem of the future betterment of the human race is confessedly, at the present time, hardly advanced beyond the stage of academic interest,  but thought and action move swiftly nowadays, and it is by no means impossible that a generation  which has witnessed the exclusion of the Chinese race from the customary privileges of settlers in  two continents, and the deportation of a Hebrew population from a large portion of a third, may  live to see other analogous acts performed under sudden socialistic pressure. (xx)

Galton here does not advocate massive social engineering. But it's clear that if the evidence supports its utility, he would embrace it. ("They compel us to face the question as to what races should be politicall the chief occupiers of that [African] continent." (xxv))

The Neuraths do not comment much on Galton's vision. So, Uebel may be right that they wished to introduce "certain sobriety" into the discussion. (As I speculated yesterday, while Galton endorses a racial hierarchy his book is not antisemitic (even philosemitic), something they may have found attractive in a Germanic context.) Perhaps they did not endorse his racial hierarchy and other eugenic projects. (But why not say so in the  introduction?) But it does not follow that they rejected everything in the book. In fact, there is clear evidence that the Neuraths explicitly accept Galton's idea of the inherited characteristic of intelligence. For, in their 'preface" (Vorwort) they explicitly call attention to a (modest) mistake by Galton. Galton had presented Fichte as an isolated genius; but they note that he had a son who was a philosophy professor. The point is echoed in the biographical sketch of Galton in their "introduction," where they explicitly note, as an illustrative example of Galton's theory, the above average likely presence of other smart people in the families of exemplary men (iii-iv).**

 

On Taboos & Polarization in Politics & Philosophy

[T]he logical empiricists can be interpreted as engaged in a project of voluntarist racial eliminativism. They did not take themselves to have an in principle argument against racial categorisation of human beings. Their frequent references to biological racial categorisation make it clear that they thought this was something you could coherently do, and it would be inconsistent with their version of conventionalism to argue that one is cognitively mistaken to use an empirically meaningful categorisation scheme. Hence, while the position I am attributing to them is similar to the position that racial taxonomic terms are simply meaningless, it is not quite that – it is not that they thought there was no empirically meaningful way that one could make sense of human racial taxonomy. Rather, it was that they thought one ought not use these terms....As such they adopted a policy of refusing to endorse any instances of racial categories being used as part of empirical explanations, and discouraging racial explanation by rhetorical means.--Liam Kofi Bright (2017) "Logical empiricists on race" Studies in History and Philosophy of Science Part C: Studies in History and Philosophy of Biological and Biomedical Sciences

Liam Kofi Bright reminded me of his paper in response to my post on the method of composite portraiture (which traces back to Galton). [I leave aside here to what degree Bright's paper fits that method.] In his paper he diagnoses a taboo within logical empiricism against (biological) racial explanations of social affairs. The taboo is maintained by an act of self-legislation in the adoption of certain conceptual frameworks (and methods) and a rhetorical-polemical attitude against those that used alternative conceptual frameworks (that included racial explanations). What makes taboos interesting is that, in order to be effective, they require self-command and group discipline/loyalty, that is, esprit des corps. This is possible when a group is reasonably small, shares an important aim, and has a way of controlling entry. Bright does not address the maintenance of the taboo, but in the sentences I skipped, Bright suggests that it is their (i.e., the logical empiricists') shared "political projects" and moral commitments, which are 'external' or 'optative' to the adopted framework that sustains the taboo (a word he does not use), undoubtedly facilitated by informal communication (when they were in the same place) and small numbers. Because Bright reads the logical empiricists in a Kantianizing fashion (with relativized a priori), we may remind ourselves that they are not the first to legislate a taboo within philosophy: in Kant's Die Metaphysik der Sitten, Kant (recall) limits -- as (ahh) Slavoj Žižek point out -- the possible philosophical role of history (as a form of inquiry) into the origin of a legitimate authority on practical grounds.

In his essay, Bright explicitly recognizes that the taboo (racial explanation) is maintained not because using it is meaningless. That is to say, the logical empiricists have not shown that using racial explanation is false. One need not be Freud to recognize one does not need taboos on the evidently false. But it also does not follow that the logical empiricists thought the taboo was hiding the truth. (If anything, using a taboo to hide the truth is self-defeating because one will generate contradictions, internal tensions, etc.) Rather, it's clear that they thought that regardless of the scientific status of racial explanation, there is always a non-trivial risk these get mingled with immoral/inhumane political aims. That is, the members of the Vienna Circle are (recall) engaged in responsible speech because they recognize what we call the inductive risks of legitimizing racial explanations. The polemic with Heidegger centers, in part, on their recognition that his way of philosophizing has noxious consequences (recall Glymour). Taboos can be efficient in preventing the predictable downstream costs and energy of combating the pernicious effects of the damage done by racial explanation (in the service of racial hierarchy, social exclusion, eugenic programs, etc.).

Bright is clear that the logical empiricists engaged in polemical rhetoric against those that violated their taboo. This is significant for two reasons: first, they attempted to impose their internal compact on others (who had not contracted into their program). That is, there is something imperialist to demand from others that they abide by your self-imposed taboo. Second, we can discern here strategies that became all too familiar in their students and heirs in later generations of analytical philosophy; an excessive willingness to use rhetorical means in combating opposing schools.

Given the political moment -- with racial subordination and elimination a live possibility -- it is easy to forgive the logical empiricists willingness to aim to impose their taboo with rhetorical means on others. They were were part of a struggle in which the racialist explanation loving sides were quite willing to support racial hierarchy, and worse. But most of the subsequent academic struggles their students engaged in are not like this; being wrong about when to discusses semantic composition or modal metaphysics does not lead to racial elimination.*

As an extensive aside, Bright recognizes a problematic exception to his argument: Neurath's translation (in 1910 with his first wife Anna Schapire-Neurath) of Galton's work on eugenics, Hereditary genius.** (Speaking of taboos, Neurath's translation goes unmentioned in the Stanford Encyclopedia entry on Neurath!) Now, Bright could have said that in 1910 logical empiricism did not exist yet and that circumstances between 1910 and, say, the 1930s were different (the founding of the Vienna Circle can be dated to 1928 or so). This would also explain Neurath's apparent change of heart later. This is not Bright's strategy (perhaps because he also quotes from the relatively early (1921) Anti-Spengler.) With a nod to Thomas Uebel's contextual analysis and, more important, the writings of W.E. Du Bois, Bright remarks that a defense of racial breeding need not be a defense of white supremacy. And, undoubtedly, there is logical space between the can wish to breed for all kinds of characteristic without wishing for a racialized polity. And, in fact, in 1910 eugenics was part of a progressive, scientific gospel that influenced all kinds of policies and scientific projects not all of them pertaining to race (see Tim Leonard's book).

Even so, it is worth noting that by 1910, Galton's Heriditary Genius, was an old book (first published in 1869); it draws on Darwin's ideas, but by 1910 was certainly not cutting edge in its science. It has, in fact, a chapter, "The Comparative Worth of Different Races" which is as bad as it sounds.*** Galton, who was the leading statistician of his age, concludes "that the average intellectual standard of the negro race is some two grades below our own [Anglo-Saxon]." (In the hierarchy, the negro is still a grade above the "Australian type.") A satirist may note that Galton is, in fact, willing to acknowledge that there were better races than his own, for he remarks that "the average ability of the [long vanished!] Athenian race is, on the lowest possible estimate, very nearly two grades higher than our own." But Galton takes this to be an argument for racial purity because the Athenians allowed too much interbreeding!+ So, Galton's Heriditary Genius clearly embraces the existence of racial hierarchies (and these are not incidentally connected to British empire). But, interestingly enough (and I suspect this may have had some appeal to young Neurath), unlike many German eugenic tracts, Heriditary Genius is not antisemitic. While the Jews are not formerly classified in the racial hierarchy (due to lack of statistical evidence), most of the references to Jews in it are favorable.

Let me return to Bright's article. He establishes the existence of a philosophical taboo in (what I'll call) a highly polarized political context. Bright does not take a stance on the strategy he describes. But I am inclined to think that sometimes such taboos are necessary and useful and that the logical empiricists are in this, and in so many other ways, worth emulating. The opposing camps were part of a larger Kulturkampf where there was a genuine risk that bullets (and worse) would settle the true about racial hierarchy [to be precise: bullets can't settle the truth]. In the context of such high degree of polarization, the voluntary enforcement of taboos can serve a valuable function. They save and harness intellectual resources, they prevent unnecessary mingling of meaningful ideas with awful political and ethical policies, and they prevent the ambitious young from being tempted down unwise and immoral paths.

Of course, the rhetorical and professional enforcement of a taboo comes with a two-fold cost. First, as we have  seen, the very success of the strategy may encourage imitation of certain, imperialist behaviors (involving less dangerous topics) in less existentially polarized contexts. Second, the existence of a taboo is also a temptation for the unscrupulous and ambitious young who find themselves born in later times. They may well think that there are low-hanging fruits which will advance the cause of knowledge (and their career). Sometimes these unwise are lucky and they live in quiet times and their greedy, rashness has no downside.

 

On Being On To Something

Professor Kenneth Arrow died on February 21, 2017, at the age of 95. He was widely regarded (along with Paul Samuelson, John Hicks and possibly --- depending on tastes --- John Maynard Keynes, Milton Friedman and Gary Becker) as one of the greatest economists of the 20th century. He also happened to be my favorite economist of all time.    

Professor Dipak Banerjee, my teacher, introduced me to Kenneth Arrow in 1974, who appeared (much in the manner of Hindu god[esses]s for whom my mother had special reverence) in the form of a small yellow paperback. I acquired Social Choice and Individual Values from Dasgupta and Co. of College Street, and still have it. I was a first year undergraduate. That little book was a repository of the most profound logical thought. I had never seen anyone distill what appeared to be an abstract question in political economy into a theoretical device that cut sharply, and cut deep.

What was the question? Briefly, it was well known from the so-called Condorcet paradox that majority voting could produce nasty cycles in choices, even when the individual preferences involved in that voting process were perfectly reasonable.  That led to the question: was there any political system that could “reasonably” aggregate individual preferences? Now think about the question for a second: we know what majority voting is, but there is in principle an infinity of other systems. How could one ever formulate such a problem, let alone attempt to answer it? The very formulation — as axioms placed on an abstract mapping that connected individual preferences to their social counterpart — was sheer genius. But the apparatus was not only beautiful: it could also speak. It argued that under the minimal desiderata placed on the aggregator, there was no way of putting together individual preferences into a satisfactory social ordering; one that was cycle-free.

Arrow had received the Nobel Prize just two years before this encounter with him. He was just 51 years old, by far the youngest Laureate then (or since) in Economics. I ran off to the National Library in Calcutta to dig out the Nobel citation, excitedly anticipating a homage to my beloved Impossibility Theorem. Yet oddly, the The Nobel citation mentions Arrow’s monumental theorem only at the very end, and almost in passing. It focused instead on Arrow’s (and Hicks’s) contributions to general equilibrium theory:

“[Arrow] provided the basis for a radical reformulation of the traditional equilibrium theory. Through this reformulation, which was based on the mathematical theory of convex sets, the general equilibrium theory gained both in generality and in simplicity… The model presented in this paper became the starting point for the major part of further research in this field. Among Arrow's many important contributions should also be mentioned his development of the theory of uncertainty and its incorporation within the frame of general equilibrium theory and, furthermore, his analysis of the possibilities for decentralized decisions in a society where the price system is fixed by the central authority… As perhaps the most important of Arrow's many contributions to welfare theory appears his ‘possibility theorem,’ according to which it is impossible to construct a social welfare function out of individual preference functions.”

This was disappointing as far as my current passion was concerned. But it was also exhilarating: because there was more! (Later, I realized just how much more.) Back I went to Professor Banerjee. I wanted to know why general equilibrium was not just a question of several equations in the same number of unknowns, and what all this was about “the mathematical theory of convex sets.” In response, Dipak-babu helpfully produced another small tome by Gerard Debreu. This was based on the work with Arrow. Though certainly more mainstream this time in its questions, the techniques went way over my 17-year old head....

As a graduate student, reading all this in late 70s and early 80s, I viewed Arrow as a thinker who could both ask the deepest questions in economic and political philosophy, and at the same  time use mathematical arguments with ease and utility to answer them to a substantive degree. This was someone who could put past accomplishments into perspective, leave them behind, and seek to look beyond to newer and more difficult concerns. Such were the nature of his forays into questions of incomplete information, pervasive externalities, increasing returns and interpersonal equity. There was just one word for it: inspirational. The inevitable reaction was not long in coming: I wanted to be like Ken Arrow. Surprise surprise, that was not destined to happen. But something else did: I had the immense good fortune to become his colleague at Stanford.

I was at Stanford on a job flyout, at the very beginning of 1982. I had been warned about the meeting with Arrow. Apparently, all I had to do was tell him the assumptions of my model and then, before I could get any further, he would proceed to tell me all the results that could conceivably be proved from those assumptions. This was unnerving news. But nothing of the sort happened.

I walked into his office. It was small and cluttered, full of books that went up high. It had a vertical rather than horizontal feel. Arrow himself gave the opposite impression. He was shorter than I had expected, strong, rooted to the ground. The man seemed to be in constant motion. He was both brisk and welcoming. He was wearing brightly colored suspenders, and there was an immense bicycle helmet on his desk. (I did not then realize that these, along with the perennially flipped pencil at seminars, would be an intrinsic part of my later memory of him.) I began talking about my work. It was a bit of an out-of-body experience; I could see myself talking to him. Arrow listened very closely. There was an intensity of gaze that never wavered, except when he would start speculating, during which he would look up at the ceiling and back to me. He asked questions non-stop. He talked very fast, the words tripping over one another, the tone uneven, the sentences clearly struggling to keep up with the flow of thoughts. He didn’t exactly anticipate my results. But after 15 minutes, I had a second eerie sensation: that I was talking to someone who  had thought about my problem for a very long time. This was a weird feeling that I came to associate with Arrow over the next few years.

....

Less enthusiasm was shown [by Arrow] towards the neo-Ricardians. Once I had mustered up enough courage, I would talk with Arrow with complete freedom, and before long I had told him that I had once spent an entire day with Piero Sraffa at Cambridge, and that he had given  me a signed copy of his book ...I told [Arrow] that I admired the book for its apparent demonstration that the distribution of income across labor income and profit could not be fully pinned down by economics — that some reference to the political system was needed. Arrow looked at me with a mix of irritation and pity: “I’ll get you out of that soon enough.” And of course he did convince me that Production of Commodities By Means of Commodities had an extra degree of freedom in it that generated a fake indeterminacy, though I still harbor a sneaking suspicion that Sraffa was on to something.--Debraj Ray "Kenneth Arrow, 1921-2017"

Of the quick, first round of obituaries of Arrow, the one by Ray, himself a leading economist (recall that I have remarked on his criticism of Piketty), stands out for its excellence mixing intellectual, psychological, and sociological observations on the significance and nature of Arrow's scholarly life as seen by professional economists today. It's worth reading in full (including the stuff on the scale and content of some of Arrow's many contributions to economics as well as Arrow's relationship to socialism) not the least for acquiring a flavor of the anecdotes that circulate around Arrow's life. One other reason why I quote so much of it -- despite some names and issues being obscure to philosophers -- is that Ray's manner of evaluation is so familiar to us professional philosophers (genius; deepest questions; out-of-body experience; intensity of gaze; talking to someone who  had thought about my problem for a very long time, etc.) They reminded me of the tenor of the recent remarks about Derek Parfit (who, to be sure, published much less than Arrow and was probably slightly less central to the disciplinary formation in philosophy than Arrow was in economics).

In the final sentence of the quoted paragraph (which is not the end of the obituary), Ray allows himself one very modest reservation: "I still harbor a sneaking suspicion that Sraffa was on to something." Now, in context, Ray is remarking on the fact that Arrow (then already very famous) "was teaching History of Economic Thought" and that among the history covered Arrow favored Cournot.* Ray is reporting on a series of exchanges with Arrow some time in the early 1980s (so more than thirty years ago). Even then Sraffa's work was not setting the agenda and conceptual framework for professional economics (in the way that Samuelson and Arrow did and to some extent still do). Even if we allow historical distortion, Arrow's irritation at the interest of his younger, talented colleague in a by then discarded project is notable.+

As an aside, as it happens in 1991,** Arrow published (inter alia) about Sraffa's role in the (Marxist leaning) reception of Ricardo, which influenced the way Keynes was assimilated, especially at Cambridge (and especially by Joan Robinson). In the paper, "Ricardo's Work as Viewed by Later Economists," Arrow self-describes himself as "a practicing theorist, not a scholar in the history of economic thought." Yet, Arrow shows great command over the historical material and the history of Ricardo reception.** One criticism of Sraffa by Arrow is that Sraffa is "indifferent to [Ricardo's] the full employment hypothesis, though one would have thought a position on this question to be basic to the formulation of a coherent model." Another (more implicit) criticism of Sraffa is that "the possibility of diminishing returns was dropped by Marx and equally by Sraffa and Joan Robinson."

To return to my main point. Ray reports his own "sneaking suspicion that Sraffa was on to something" despite being exposed to (presumably conclusive) proof that Sraffa's attempted formal "demonstration that the distribution of income across labor income and profit could not be fully pinned down by economics" was incoherent or unsupported. Lurking in the background is not just the question of the role of politics in determining economic outcomes, but also the question of disciplinary autonomy and closure. One would have liked to hear more about Ray's suspicions here -- Ray is the editor of the American Economic Review so he plays a non-trivial role (among many others) as gate-keeper and standard setter in the profession.++ 

But leaving aside the question of Arrow's criticism of Sraffa, I think the phenomena that Ray puts his finger on is more widespread (and also prevalent in philosophy): often the technically superior person will refute a theory Y -- and let's stipulate these are not straw-man versions of Y --, and [this matters] will do so in ways such that new recruits will not drift toward Y or variants of Y, and yet the insight (u) that gave rise to Y may not have been adequately handled.*** This is due to the fact that the relationship between (u) and Y is not a tight mapping. Much formalization in economics (recall my piece on the criticism of Harberger by MA. Khan) and all explication (recall my post on Leitgeb, Carnap, and Tarski) have some degrees of freedom and judgment about the way in which a would be Y captures (u), especially if (u) is interesting. After the refutation of Y, the partisan of (u) is often told to keep quiet until she can offer a successor to Y  that is not prone to the fatal flaws exposed of Y (by technically superior person). [This is even so in circumstances where Y could handle stuff in ways that no other present theory does, as in cases of Kuhn loss (recall).]

If (u) is something non-trivial, even important, it can be problematic that a discipline avoids working on developing successors or replacements of Y. And so, ideally, one would wish that the refutation of Y would be accompanied by some theoretical articulation that does justice to the significance of (u). [Notice that this is true of formal and non-formal theories of Y.] But currently the incentives and norms of many disciplines work against this and we are not encouraged to reflect on how this might be improved.

 

On Over-Reliance on Bayesianism in Macro-economics

The identification problem means that to get results, an econometrician has tovfeed in something other than data on the variables in the simultaneous system. Ivwill refer to things that get fed in as facts with unknown truth value (FWUTV) to emphasize that although the estimation process treats the FWUTV’s as if they were facts known to be true, the process of estimating the model reveals nothing about the actual truth value. The current practice in DSGE econometrics is feed in some FWUTV’s by "calibrating" the values of some parameters and to feed in others tight Bayesian priors. As Olivier Blanchard (2016) observes with his typical understatement, "in many cases, the justification for the tight prior is weak at best, and what is estimated reflects more the prior of the researcher than the likelihood function."

This is more problematic than it sounds. The prior specified for one parameter can have a decisive influence on the results for others. This means that the econometrician can search for priors on seemingly unimportant parameters.--Paul Romer (2016) "The Trouble With Macroeconomics." [HT: LOTS OF PEOPLE]

Two, relatively recent methodological pieces by Paul Romer, an influential economist, have received a lot of media attention. The one quoted above and another (about mathiness). "The Trouble with Macroeconomics" argues that Macroeconomics has regressed "into pseudoscience." It partially explains this regress in terms of the sociology of the discipline with (i) too much "respect for highly regarded leaders evolves into a deference to authority," and (ii) a research culture that values certain kind of technical  sophistication (e.g. 20). This part of the paper made a splash because it did not pull punches and named names. I am going to ignore this side of the paper, although it reminded me a lot of bits of professional philosophy at various times. (Romer explicitly compares the state of affairs with the situation in String Theory as presented by Lee Smolin.)* Here I focus on the more technical and evidential criticisms by Romer of macro-economics.

The key technical issues are related to an issue that is known as the 'identification problem.' To put the point informally and in very traditional terms (familiar from the long history of science before the development of econometrics): observation (data) cannot decide among observationally equivalent models (even with the very best statistical techniques). The research community cannot figure out the parameters of an equation or model that track causes in the world (or the model-world). This is especially difficult when, as Romer notes, variables are part of a simultaneous system. And "when the number of variables in a model increases, the identification problem gets much worse." (6) So far so good.

Now, leaving aside the sociological issues  noted above, Romer identifies two further problematic features of the research practices of modeling community: (iii) the most sophisticated models introduce what Romer calls "imaginary forcing variables" that (a) do not refer to any underlying economic behaviors (or emergent properties of such behavior) -- a kind of fudge factor -- and that (b) are used to make sure that the model gives unique answers (and so -- Romer does not mention this -- can be used in a policy context). These imaginary forcing variables are treated in satirical fashion by Romer and also allow Romer to refer to phlogiston. This is actually very unfair to phlogiston theory because phlogiston was, in fact, taken to be a real entity that could be measured and explained observations (and in some conceptual schemes was measured, but leave that aside).

And, as the quoted passage above notes, (iv) the abuse of Bayesianism, which makes it too easy for scholars to produce results. One problem here is that the priors are not being set by well-confirmed background theory (there is none, so the likelihood function is pretty much meaningless), but by the interests of the researcher to produce results (for career advancement, policy status, etc.). This kind of abuse of Bayesianism is entirely foreseeable, and one wishes that philosophers of science, who as a community have grown ever more enchanted with Bayesianism, would recognize that in application Bayesianism is liable to a certain sort of systematic abuse and, recall, expert overconfidence. (There are actually multiple abuses here: fake precision, too much permissiveness, and it focuses folk on confirmation rather than on stress-testing the concepts). This is not to deny the utility of Bayesianism as one of many research instruments, but its advocates tend to ignore the down-side risks of adopting it in non-idealized scientific communities.

Both (iii-iv) are connected to and facilitate a more general problem with the research culture in macroeconomics. There is a general bias toward confirming models than in stress-testing the particular variables/concepts that enter into the model (something I have been saying for over a decade now [see also this piece by Abe Stone for more general point]). With the demise of Popper's philosophy of science, it's been very difficult to get people to see that stress-testing of concepts and measures is important. How to design theory and models that makes such stress testing possible and to turn date into high quality evidence about parameters is for another time (although I warmly recommend folk to read George Smith's work on Isaac Newton, see, e.g., Closing the Loop.)**

 

On Stasavage, Hume, and Machiavelli: Credit, a proto-Geneology of the Origin of Colonialism and the State as Enterprise

In 1339 under Genoa's first doge...a new principal [sic?] was established of paying variable interest on the consolidated debt, depending on the level of customs revenue, which in turn depended on the level of trade. As a result of this change, holding Genoese debt began to resemble holding an equity stake in this state. David Stasavage (2011) States of Credit: Size, Power, and the Development of European Polities, 118. [HT Deborah Boucoyannis]

The passage I quote from Stasavage is an aside in the larger, fascinating argument of the book. The book's main argument is about the relationships among territorial size, the nature of political representation, monitoring costs, and the origin of state debts. His core conclusion is that in Europe (in a band originating in Carolingian Lotharingia) merchant oligarchies in relatively compact territories and with (modestly) representative institutions were able to solve all the monitoring and principal agent problems such that states could establish long term debt on relatively cheap terms. And this allowed these city-states to remain independent (or in the case of the Dutch Republic become independent) in the face of much larger and more populated state rivals. Methodologically, the book is interesting because it combines econometric techniques, historical research, and qualitative research to offer causal history that takes the interactions among geography, political economy, and political history seriously.*

Stasavage is not the first major political theorist to call attention to the significance of Genoese political economy for political theory. Inspired by Machiavelli, Hume had done so in his  clever and  important (1742) essay Politics May Be Reduced to a Science. Interestingly enough Stasavage mentions and discusses Hume's source, Machiavelli's Florentine Histories (see pp. 117-122), but not Hume's analysis. Stasavage treats Machiavelli's account (it's in Book 8, chapter 6 of the Florentine History) as a rival to his own. Stasavage interprets Machiavelli as claiming that with the establishment of the Casa di San Giorgio a governance problem was solved, by insulating financial management from political turmoil, and therefore interest rates dropped. But Stasavage notes that interest rates dropped much further and more durably only after the Andrea Doria Republic was establishment (that is, Genoa became a permanent oligarchic polity).

While I wouldn't say that Stasavage is unfair to Machiavelli (who is really not interested in interest rates), he kind of misses Machiavelli's point. Hume's analysis is much closer to  the mark: "And so little dependance has this affair on the humours and education of particular men, that one part of the same republic may be wisely conducted, and another weakly, by the very same men, merely on account of the difference of the forms and institutions, by which these parts are regulated." Hume then gives Casa di San Giorgio and Genoa as an example, and then cites and quotes Machiavelli in Italian as his source. That is, Hume treats the example as an exemplar of the significance of institutional design and the role of incentives in them (in part to rule out competing explanations of differences in political governance). In this way Hume already anticipates part of Stasavage's own instititutional insights.

As an aside, Machiavelli's own methodological conclusion of Casa di San Giorgio case is to treat it as "A truly rare example, and one never found by the philosophers in all their imagined or dreamed of republics, to see in the same circle, among the same citizens, liberty and tyranny, the civil and the corrupt life, justice and license." That is, rather than opting for the method of thought experiments (and ideal theory), philosophers should use actual history and what we would call 'empirical social science' as the evidential input for their theorizing because (a particular) philosophical imagination tends to be more impoverished than reality. (This too inspired Hume and, earlier, Spinoza (recall).)

Even so, the passage I quoted from Stasavage above calls attention to a remarkable fact that once variable interest dependent not just on taxing power but also on underlying economic conditions was introduced, state creditors are turned from citizens into something more akin to shareholders of a corporation. When these shareholders are also the ruling elite they run, as Stasavage shows repeatedly, the corporation for their own economic interests by, for example, pushing the tax burden on (regressive) excise taxes (which hit the consumption of the working poor especially hard). In Stasavage's account this is also a source of political instability because it generates popular revolts.

The idea of equity in a state is a significant one if we peek ahead at the history of imperialism. We tend to forget that the great Dutch and British colonial enterprises were originally built around corporations (state granted monopolies with trading privileges), their East India [and West India] Companies, who treated the people and land they acquired as property of the corporation to be used for trade and production. Shares of these corporations were listed and trade-able, paying out dividends. In the Dutch case, there was, in fact, great overlap between the political elite and merchants/shareholders who ran the trading companies. (I am less familiar with the British case.) This is what led Adam Smith to claim that "The Government of an exclusive company of merchants is, perhaps the worst of all governments for any country whatever." In context he is talking of the Danish control of St. Thomas and Santa Cruz, but from larger context it's clear he is describing all colonial Corporations. But since Smith was intimately familiar with Hume's economic writings (and well read in Machiavelli), and since he emphasizes 'any country whatever' it is not impossible that he is also thinking of (Genoa's) merchant olicharchy.

I close with two observation. There is sometimes a tacit tendency to think that the increasing focus on economic parameters in analysis of politics is a consequence of the dominance of economic theory. This is often accompanied by the thought that economics is the only or the major ruling value. Often the thought is followed by a lament of the loss of influence of other values (solidarity, community, etc.). And, indeed, if we look at the passage quoted from Machiavelli we see how list of values ("liberty and tyranny, the civil and the corrupt life, justice and license") that do not enter into technocratic policy science (with possible exception of justice). So, my first observation is that the recent U.S. Presidential election reminds us of the robustness of Machiavelli's list. Second, that indeed sometimes practice precedes theory/thought greatly. The Genoese invented oligarchic, state capitalism long before it was exported as imperialism outside of Europe and theory caught up with the possibility.

 

 

"Old Ideas in Pure Theory," On the Pre-History of Performativity and Reflexivity in the Theory of Markets under Uncertainty

The Arrow-Debreu theory was not originally put forward for the case of uncertainty, but an ingenious device introduced by Arrow, and further elaborated by Debreu, enabled the theory to be reinterpreted to cover the case of uncertainty about the availability of resources and about consumption and production possibilities...[Arrow's device permits one to incorporate] uncertainty about the environment into a Walrasian model of competitive equilibrium. The basic idea is that commodities are to be distinguished, not only by their physical characteristics and by the location and dates of their availability and/or use, but also by the environmental event in which they are made available and/or used. For example, ice cream made available (at a particular location on a particular date) if the weather is hot may be considered to be a different commodity from the same kind of ice cream made available (at the same location and date) if the weather is cold. We are thus led to consider a list of "commodities" that is greatly expanded by comparison with the corresponding case of certainty about the environment. The standard arguments of the theory of competitive equilibrium, applied to an economy with this expanded list of commodities, then require that we envisage a "price" for each commodity in the list, or, more precisely, a set of price ratios specifying the rate of exchange between each pair of commodities.

it will be useful to give a more precise account of the concepts of environment and event that I shall be employing. The description of the "physical world" is decomposed into three sets of variables: (1) decision variables, which are controlled (chosen) by economic agents; (2) environmental variables, which are not controlled by any economic agent; and (3) all other variables, which are completely determined (possibly jointly) by decisions and environmental variables. A state of the environment is a complete specification (history) of the environmental variables from the beginning to the end of the economic system in question. An event is a set of states...Granting that we cannot know the future with certainty, at any given date, there will be a family of elementary observable (knowable) events, which can be represented by a partition of the set of all possible states (histories) into a family of mutually exclusive subsets. It is natural to assume that the partitions corresponding to successive dates are successively finer, which represents the accumulation of information about the environment.

The definition of equilibrium requires that the agents have the access to the complete system of prices when choosing their plans. In effect, this requires that at the beginning of time all agents have available a (common) forecast of the equilibrium spot prices that will prevail at every future date and event. Radner (1970) "New Ideas in Pure Theory: Problems in the Theory of Markets under Uncertainty," 454-6 [HT M. Ali Khan]

The title of this post is an allusion to Radner's comment (454) that he had intended to submit his paper to a session called, "Old Ideas in Pure Theory." Radner treats so-called, Arrow or Contingent Commodities (recall) as a way to handle "uncertainty about the environment" within a Walrasian model of competitive or general equilibrium. Arrow's key (1953) insight is to connect such a commodity to every possible state (date and event) of the environment in advance. This commodity is a kind of insurance (or option). Below I return to this move as a way to handle uncertainty. One assumption that Arrow (and Debreu after him) relies on is that "at any one date each agent will have incomplete information about the state of the environment, but all the agents will have the same information." That is to say, Arrow had been working with a kind of representative agent model. One of Radner's contributions to this literature is to show how to represent "differences in information among the economic agents." (456)

Now, while extending the model, Radner also recognized several limitations on the Arrow-Debreu model. One line of limitation is that "the theory does not take account of at least three important institutional features of modern capitalist economies: money, the stock market, and active markets at every date." (457) And, in fact, Radner goes on to suggest quite elegantly that money, the stock market, and active markets are responses to uncertainty not just about the environment but also uncertainty about future prices. (457-8) To put the significance of this point poetically -- and, perhaps, go against Radner's  more sober spirit -- here Radner shows how from within a pure (friction-less) model universe, the development of capitalist institutions that allow agents to manage uncertainty may be understood without formalizing this understanding. That's not the end of the matter because Radner then goes on to summarize his earlier results (now known as a Radner Equilibrium) what active (or "incomplete") markets may look like in his model universe (and so make the Arrow-Debreu far more dynamic):

Suppose that no market at any one date is complete in the Arrow-Debreu sense; i.e., at every date and for every commodity there will be some future dates and some events at those future dates for which it will not be possible to make current contracts for future delivery contingent on those events. In such a model, several types of "equilibrium" concept suggest themselves. First, we may think of a sequence of "momentary" equilibria in which the current market is cleared at each date. The prices at which the current market is cleared at any one date will depend upon (among other things) the expectations that the agents hold concerning prices in future markets (to be distinguished from future prices on the current market!). We can represent a given agent's expectations in a precise manner as a function (schedule) that indicates what the prices will be at a given future date in each elementary event at that date. This includes, in particular, the representation of future prices as random variables, if we admit that the uncertainty of the agent about future events can be scaled in terms of subjective probabilities [16].

Radner's reference ("[16]") is to Savage, who had indeed represented uncertainty about future events in terms of subjective probabilities, but the more fundamental trick -- the representation of future prices as random variables --, is (recall) due to Alchian (and flagged by Arrow), but by 1970 was being used more widely (recall Foucault on Becker 1962). So, in Radner's paper we find the three standard ways in which true Knightian uncertainty is displaced by three successor concepts in pure theory (or mathematical economics): (i) treat it as subjective probability (Savage); (ii) treat it as Randomness (Alchian); (iii) treat it as an Arrow commodity.* (It's possible that Radner is the first to use them jointly.) Elsewhere I have explained (for example here) why (i-iii) fail to do justice to Knightian uncertainty (and I add a note below).

Now, the resulting model has one interesting feature that is flagged by Radner: "since the equilibrium price expectations are self-fulfilling, the observation of the prices in any current market provides information about the true state of the environment (i.e., the specification of the values of particular prices defines an "event" in the set of possible states of the environment)." (459) To the best of my knowledge, Radner is the first formal modeler to note this species of performativity in a (model) market economy.

Some other time I explore what happens if one confuses this model for reality. I close with another observation. Radner had noted another limitation of the original Arrow-Debreu model:

If the Arrow-Debreu model is given a literal interpretation, then it clearly requires that the economic agents possess capabilities of imagination and calculation that exceed reality by many orders of magnitude. Related to this is the observation that the theory requires in principle a complete system of insurance and futures markets, which appears to be too complex, detailed, and refined to have practical significance. (457)

Without noticing  it, perhaps, Radner here echoes one of Keynes's (1921) original reasons for embracing the reality of uncertainty. The cognitive capacities of economic agents are limited. The key word, in fact, is "imagination." After all, cheapening computer power is quickly diminishing the problem of "calculation." Rather, in the Arrow-Debreu universe we need to imagine future possible states exhaustively in advance. Elsewhere, I have complained about the modal commitment to finite, countable universes built into this model (which is supposed to model uncertainty); but even if we leave aside modal metaphysics, Radner is right to note that it is peculiar to think one can eliminate true uncertainty by relying on the imaginations of economic agents; if agents were imagining the unknown unknowns, they wouldn't be.