Responses to "The Designer Economy" by Yakov Feygin and Nils Gilman

trombley talks philosophy and alex talks implementation

and

Jan 25, 2023

Alex’s Response

Heterodox economists want a changeover to their viewpoint to go down like a political revolution: going from no one acknowledging them as the experts to everyone acknowledging them. This misunderstands how revolutions work. They are almost never grand confrontations, but rather situations where one organizing principle (embodied in some group) falters sufficiently that a different already-ongoing organizing principle (embodied in some other group) is able to successfully assert itself as “primary.” Sometimes grand confrontations happen in the course of that, but more often, they look like questions of the subsumption of administrative behavior under competitive rule-makers.

I don’t think they’re going to get their preferred revolution, but I do think that they can accomplish the bulk of its substantive aims in different ways. An alternative way of thinking about this is probably Deleuze’s “enculage,” which I will not translate here, other than to say that it ends in an apostrophe rather than a “g”. Judo is probably more family-friendly: redirecting your opponents’ efforts such that although they may exercise more local force, the global arrangement of forces shows the agent exercising less force to “win.”

I think the Designer Economy frame is a good way to think about how to judo certain existing literatures within economics. In particular, Mechanism Design. People got really excited about this in the 1990s, when it looked like the “end” of History was the end of history. Instead of statecraft, we would just have to “design auctions appropriately” or “design markets,” and then the rest of the virtues would necessarily follow.

What this literature conceded at the time that was kindof radical was the idea that markets could be “designed” rather than deferred to. The problem was, their vision of Good Market Design was based on mid-century Chicago Price Theory. The goal was to Eliminate Harberger Triangles Through Magic Or Science, and we saw them try this in things like the government EM Band auctions in the 90s and 00s. The extreme end of this today is Glen Weyl and Richard Posner’s kid arguing that we should just make one huge database so that Americans can spend their lives snipe-trading one anothers consumer goods.

What the Designer Economy piece implies is that mechanism design as practiced in the past is good, but that the control parameters for markets should not be prices, because prices are basically ephemeral. Instead, we most probably want to look at statistical distributions of prices: volatility, skew, whatever higher moments matter for participants in a given market. That is because this embeds the role of Marshall’s “well-informed dealers” in ways that satisfy the now-retro demand that markets “act as information processors.” No one producing anything worthwhile [ex-Finance] is looking at exact sequential prices to make the kinds of production and investment decisions that Austrians and Chicago Types believe are downstream of the distribution of relative prices. Instead, they are looking at the distributions of realized prices in order to do fuzzy logic on the trends of a given input or output.

Think about the ways in which everyone freaked out about high oil prices. There was a signal there (please use less gas while we figure this out), but it did not need the full extent of the realized volatility to pass that signal. One of the things I’ve always meant to write out is a model I have in the back of my head of technical change as convexity hedging against expected wage movements. Oh well, not today.

The trouble with parameterizing on higher moments in the design of specific markets is that it requires actually really knowing those markets. This leads to a question of “regulatory capture”: if we assume that everyone who knows an industry well enough to design a market for them, how can we ever get markets designed with the public purpose in view?

This is the crux of the political question here: none of the five groups that Feygin and Gilman profile as the main agents in the Designer Economy necessarily have a clean handle on how to do that exactly. It is the job of the economist – in its proper sense as “student of the economy” – to show them how.

This is why I think we need to radically refactor the Mechanism Design literature, because it could provide an actual Field Of Play for those main agents to argue for different policy priorities while sharing a space of economic possibility. This vision of economics sees it as a tool for ruling out impossible outcomes, rather than making tradeoffs. The point of governing markets in this way is to make certain undesired outcomes “impossible.” This is really different from the predict-correct-predict model of policy analysis which people like Robert Lucas (ugh) showed the flaws of. It’s also different from “planning” in most folks’ conceptions.

If someone would go forth and conquer this, I would really appreciate it. I have to tell the Fed to stop hiking right now.

Trombley’s Response

You might say it is unusual for me to talk about living people on CVAR, but our friends Yakov Feygin & Nils Gilman (F&G) recently wrote an article very much worth reading titled “The Designer Economy”. I think this article has a lot of important things to say, so let’s go over it.

The first part of the article discusses the macroeconomics of Thomas Piketty. Piketty’s theory isn’t 100% clear - he was taken to task for lazy theorizing in Romer’s Mathiness article - but I will give a quick Piketty-like theory which I think will do the bill. The world is divided into two classes, blue collar and white collar (you can say ‘global north’ and ‘equatorial’ if you prefer). The white collars hold financial capital and the blue collars earn wages, and so effectively bet on physical capital. Finance pays interest, but wages appreciate due to labor saving technical growth. Pikketty assumes (in the text, but not mathematically) that technical growth has a sigmoidal shape. Piketty further argues that financial capital in the long run has a higher rate of return than wage growth. I understand this as the liquidity premium on finance capital over physical capital, with the understanding that if this liquidity premium goes negative the whole system blows up taking wages with them. Anyway, the result is that if the white collars are allowed to collect their wealth, the sigmoidal wage growth and exponential interest growth work together so they become exponentially wealthier than the blue collars in the long run.

The upshot to this story is that a market society is stable at low growth and high income inequality. There are issues with the story. One is historical: there are many episodes where income inequality declines. Pikkety argues these are non-market events - mostly wars - and spends the majority of his book interpreting them. I will not critique this approach to market laws and history, because it is well discussed in Kindleberger’s Economic Laws And Economic History.

The other is the crucial low growth assumption. If labor saving technical growth comes fast enough and markets are in backwardation, then there is a premium for owning physical capital. The rich can make money investing in plant, making the above distinction between blue and white tricky. My understanding is that high rates of labor saving technical growth and long periods of backwardation typified the 20th century.

What F&G offer is an alternative mode of governance to the governance structure underlying Piketty’s pessimistic model. The alternative offered is not planning in the traditional sense of growing x bushels of corn in order to feed y tons of pork, but rather planning in the sense of design: creating new opportunities.

The example to follow is the COVID19 vaccines, where the global university/pharmaceutical complex with all its sclerosis managed to create a highly effective vaccine extremely rapidly. The technique used to break through the sclerosis was world governments guaranteeing demand for the product, which created enough of a pull to overcome the liquidity preference of those institutions.

The proposal of F&G is ‘designer’ in another sense as well. The Designer Economy aims to work for “every one” (“each”) rather than “everyone” (“all”). I here use the singular noun “everyone” or “all” to refer to the single mereological sum of the citizenry. By distinction, “every one”/ “each” plays the usual quantificational role.

The special grammatical nature of “everything” and “something” as distinct from simple quantification was developed by Wilfrid Sellars in his book Naturalism And Ontology, the Dewey Lectures of 74. He doesn't tie those words to mereology, at least not there.

For now, let’s get back to F&G. It may help to illustrate the way the “Designer Economy” works for each by way of contrast with a proposal that works for all. The classic proposal for a “society for everyone” is the classical utilitarian constitution developed piecemeal by Jeremy Bentham. Bentham was inspired by then novel Eulerian concepts of negative number and function. Bentham was, in fact, born at the same time Euler published the Introducito and was highly familiar with Euler’s work in fluid mechanics - see his 1778 letter to Rev. Forster where Euler is pointed out as the only “German” whose opinions on shipbuilding are worth anything and 1779 letter to his brother where urges him to read Euler - but I digress. Much like how Euler analyzed the motion of the moon as periodic signal into basic terms, Bentham proposed the basic motions of society could be synthesized into an overall signal. Again, just as the total motion of the moon is the object of real interest in Euler’s lunar theory, the total signal is the object of interest in Bentham’s social theory.

The Designer Economy is aimed at ‘each’ rather than ‘all’. Of course, the designer economy doesn’t necessarily conflict with utilitarianism, indeed it’s possible that the only way to implement a society best for all is to make a society built for each. Still, it might be worth thinking about Kantian approaches to ethics with respect to constitutions.

An economy that aims at ‘each’ rather than ‘all’ does not mean that there will be no political economy. Rather, F&G predict that if the Designer Economy continues to grow, there will be a return to more traditional political conflict over the spoils of society rather than the ‘vetocracy’ brinkmanship currently seen.

All this talk of Bentham and Kant is a bit abstract. Perhaps a more concrete example is due. The Designer Economy ‘society built for each’ concept has implications for the international finance system. Books like Eichengreen’s Golden Fetters have emphasized that the classical gold standard was a system that required much management. This is not mere hindsight, as Schumpeter develops in History Of Economic Analysis, the gold standard was normal as in normative than normal as in frequent. The gold standard was a payment system designed to blow up against deviations from the classical liberal ideal and even threats of such deviations. Schumpeter called the gold standard freedom “in the bourgeois sense”. The gold standard created an international equilibrium by delimiting the choices of actors so much that they had no ability to conflict.

Yes, the gold standard allowed central banks to act as a quasi-vetocracy constraining policy space. Athwart this classical liberal “bourgeois freedom”, we have the option of an embedded liberalism’s notion of freedom: “using discipline to enhance policy space”. The goal is to create international equilibrium by allowing states to do what they need without structural risk to neighboring states. Whether we call this an Embedded Liberalism or the Designer Economy, the new goal remains: to expand the virtual.

Okay, where did that talk of The Virtual come from? Don’t worry, this isn’t going to become an essay on why Terence Malick’s later films are actually so good (though some of them really are better than you all acted!).

The notion of the virtual was abducted from John Duns Scotus and classical mechanics. Following Doctor Duns, Peirce defined virtual as a an adjective as follows “a ‘virtual x’, where x stands for a common noun, means something which is not an x, but which has, for whatever purpose may be uppermost, the virtue of an x, that is, such properties as make it equivalent to an x.”. One can also use an adverbial form ‘virtually y x’ - where y is an adjective and x is a common noun - in the obvious way.

I generally use the term ‘virtual’ slightly differently in that I do not require a ‘virtual x’ to not be an ‘x’. This is to bring the use of ‘virtual’ in line with the use of ‘possible’, as we do include the actual under the heading possible. This makes it easier to adopt Leibniz’s stance in Theodiceé and elsewhere that the virtual is the limits of possibility, which is a key distinction when talking about a theory of the expansion of the possible. For instance, to talk of a point of a pendulum swing being the “lowest possible” is a roundabout way of discussing the actual constraints, which are virtually forces on the bob anyway.

A ball bearing is a virtual sphere because the ball bearing and the sphere both share the property ‘rolls’. Sphericity is therefore a relatively concrete predicate which can be abstracted to Virtual Sphericity and applied more widely. Doctor Duns’ most famous thesis with respect to the virtual is his conclusion that existence is the most abstract predicate. That is to say, nothing virtually exists except what really exists. Duns dubbed this the ‘univocity’ (single meaningness) of being. Kant, perhaps more shrewdly, concluded that being wasn’t a predicate at all.

All this is to say that ‘expanding the virtual’ is creating actual things which may not really be the precise object of interest. But what are the actual things created that makes the Designer Economy work.

Again the example of the COVID19 vaccine is instructive. What world governments did to rapidly create an effective and safe vaccine was guarantee demand for them. They made and kept credible promises. Guarantees, funds and so forth are the actual objects which are virtually other objects.

The general term for the “so forth” above is liquidity. As economist Maynard Keynes expounded, the very meaning of a ‘monetary economy’ is that production is aimed at liquidity rather than use.

The gold standard is an attempt to solve this problem by refusing to create liquidity, making the meager temptations of use the only option. At heart, the Designer Economy is an attempt to solve this problem by refusing to not create liquidity. Obviously I expect that - like the gold standard era - the Designer Economy will have times when it is more ideal than actual. But if I appreciate F&G’s analysis, it is because I think The Designer Economy is a better ideal than at least this alternative.