Krugman traces the idea masquerading as theory (“Modern Monetary Theory”) to Abba Lerner’s “functional finance” doctrine from 1943:

His argument was that countries that (a) rely on fiat money they control and (b) don’t borrow in someone else’s currency don’t face any debt constraints, because they can always print money to service their debt. What they face, instead, is an inflation constraint: too much fiscal stimulus will cause an overheating economy. So their budget policies should be entirely focused on getting the level of aggregate demand right: the budget deficit should be big enough to produce full employment, but no so big as to produce inflationary overheating.

Simply put, the idea is that sovereigns with obligations in their own fiat currency are constrained only by inflation in how much debt they can pile up. Krugman points to the potential problem of snowballing debt whereby debt servicing claims a larger and larger portion of the public purse. But this is a function of interest rates:

If r<g, which is true now and has mostly been true in the past, the level of debt really isn’t too much of an issue. But if r>g you do have the possibility of a debt snowball: the higher the ratio of debt to GDP the faster, other things equal, that ratio will grow. And debt can’t go to infinity — it can’t exceed total wealth, and in fact as debt gets ever higher people will demand ever-increasing returns to hold it.

So here we have another effective constraint besides inflation. How much debt sovereigns may pile up is a function of the compensation demanded by investors. Now this compensation is not uniform across borrowers. Far from it. Figure 1 displays spreads against the German bund for selected sovereigns.

Figure 1. Selected sovereign bond spreads.

The United States is much further along in the monetary cycle than the eurozone. But why do Italy and Portugal have to pay so much higher to access capital markets than Germany? The next figure shows that there is no relationship between debt-to-GDP ratios and sovereign bond yields. The rank correlation coefficient is not only insignificant but bears the wrong sign (r=-0.15, p=0.40). Restricting the sample to so-called emerging markets does not affect the result (r=-0.04, p=0.82). Note that we have already excluded Argentina (debt ratio of 57% and bond yields at 26%) and Japan (debt ratio 253% and yield 0%) since they are clear outliers. So there is simply no evidence than bond markets pay much attention to the debt burden of sovereigns.

Figure 2. The data is from 2017.

So how do bond markets judge sovereign borrowers? The short answer is that yields compensate bondholders for a number of perceived risks. Sovereigns may default, inflation may erode the value of the bond, exchange rate movements may impose losses on bondholders, interest rates may rise and thereby reduce the value of their bond. Moreover, bond yields reflect compensation for not just the expected value of the bond but also for the risk that the value may deteriorate, if for no other reason than that markets can be fickle (so that tomorrow you may not be able to sell your bond for the price you paid for it even if the price you paid was considered by all to be fair today). All these risks are constantly reevaluated by markets. The diachronic pattern is controlled by the market price of risk, itself a function of risk appetite in global markets. The synchronic pattern on the other hand is controlled by the status of sovereigns.

Some sovereigns are regarded by bond markets as safe asset providers. I have identified some safe-asset providers in Figure 2. The main one missing is Japan which would be far out to the right and bottom. Because the sovereign debt of safe asset providers is perceived to be credit default-remote and relatively protected against inflation and exchange rate movements, these assets can serve as collateral in the wholesale funding flywheel, the core of global financial intermediation. It is the practices of institutional players in this ecosystem that determines who is and who is not a safe asset provider. Safe assets can be identified by what happens to yields when the market as a whole tanks. The diagnostic pattern is that when shit hits the ceiling safe assets go up in value as investors flee to safety.

Hélène Rey has identified the curse of the regional safe asset providers. These are small countries whose debt is regarded as safe in wholesale banking practice. Even if their central banks would like to push up yields (say to defend their currency or fight inflation), adverse market developments may send them tumbling down. This is what happened to the Swiss central bank in 2015. Regional safe asset providers

… face a variant of the old ‘Triffin dilemma’: faced with a surge in the demand for their (safe) assets, regional safe asset providers must choose between increasing their external exposure, or letting their currency appreciate. In the former case, the increased exposure can generate potentially large valuation losses in the event of a global crisis…. In the limit, as the exposure grows, it could even threaten the fiscal capacity of the regional safe asset provider, or the loss absorbing capacity of its central bank, leading to a run equilibrium. Alternatively, a regional safe asset provider may choose to limit its exposure, i.e. the supply of its safe assets. The surge in demand then translates into an appreciation of the domestic currency which may adversely impact the real economy, especially the tradable sector. The smaller the  regional safe asset provider is, the less palatable either of these alternatives is likely to be, a phenomenon we dub the ‘curse of the regional safe asset provider.’

Large safe asset providers on the other hand are not so cursed precisely because of their size. But the more important point for our purposes is that big safe asset providers (Germany, Japan, and above all, the United States) are not, and in fact, cannot, be punished by bond markets for fiscal profligacy. The reason for that is the structural shortage of safe assets in the global financial system.

Figure 3. The consumption-to-wealth ratio, a measure of the financial cycle, predicts global real rates. Both are computed as US-UK wealth-weighted averages. Source: Farooqui (2016).

The issue is not whether “MMT” holds in some toy model. The issue is to what degree sovereigns are disciplined by the bond market. The answer to that question depends on their structural position that is in turn determined by wholesale banking practice. The big safe asset providers — the United States, Germany, and Japan — face considerable slack in bond market discipline because the world can’t get enough of their debt. Sovereigns not thus privileged are more exposed to bond market discipline. They may indeed have to worry about market perceptions of their public finances.

With inflation still pretty much dead and policy rates pretty much still on the floor, there is simply no case to be made for fiscal discipline for the big safe asset providers. In effect, big safe asset providers are not debt constrained. And this state of affairs will continue until the global financial system is transformed beyond recognition. It is in this sense I believe that Adam Tooze champions “MMT.”

I am not suggesting that President Warren should go on a debt binge. But worrying about bond market discipline for fiscal profligacy is to worry about precisely the wrong problem. The United States can afford to double its outstanding debt-to-GDP ratio from 100 to 200 percent and will still be perceived as less risky than Japan.

Excerpt from Rob DeSalle and Ian Tattersall’s excellent Troublesome Science: The Misuse of Genetics in Understanding Race (2018).

Linnaeus was entirely correct in assuming we should know who we are (and, even more importantly, who we want to reproduce with). But complications have arisen anyway. In the interests of practicality, modern taxonomists have amplified the concept of the type specimen far beyond that of the holotypes from which original descriptions are drawn. “Allotypes,” “neotypes,” “syntypes,” and “lectotypes,” among many others, can be invoked when a researcher identifying a type encounters some procedural hitch. For instance, when a holotype is lost (which happens occasionally, because of bombing, bad curation of a collection, or other causes), there are rules to govern its replacement by a lectotype, which will now be the “go-to” specimen.

Since Linnaeus didn’t designate one, there is and never was a holotype for Homo sapiens. Complicating matters is the fact that in the definitive tenth edition of his great work Linnaeus also described the six variants of Homo sapiens we mentioned earlier—Ferus, Americanus, Europaeus, Asiaticus, Afer (African), and Monstrosus. The first and last can be discarded as valid subspecies names, as they do not describe real specimens (Ferus was used to designate feral children, and Monstrosus was used to denote mythical people with strange morphologies.)

Under current taxonomic rules, there is one subspecies missing here, as a result of applying the principle of coordination (article 43 of the ICZN). This states that one subspecies of any subdivided species must bear the species name. What this means is that, if we are to recognize subspecies within Homo sapiens, we need to add the subspecies Homo sapiens sapiens—logically the subspecies Linnaeus had in mind when he wrote his description—to the other three.

In 1959 William Stearns, a taxonomist writing about Linnaeus’s legacy, suggested that Linnaeus himself should be the type specimen for Homo sapiens sapiens (the name that under the new rules automatically replaces H. s. europaeus). And, given Linnaeus’s estimable opinion of himself, it is certainly not out of the question that he had himself in mind when he gave our species its name. Since we don’t know for sure who he had in mind as exemplar, though, he would have to be designated as a lectotype to satisfy Stearns’s proposal. Of course, this would not help very much, because Linnaeus currently reposes in a churchyard in Uppsala, and it is not really practical for any contemporary taxonomist actually to use him as a standard of comparison.

Availability is a key consideration, and technically it became a factor some thirty years after Stearns made his original suggestion, when a group of researchers hoping to honor Edward Drinker Cope (and apparently unaware of Stearns’s proposal) suggested that Cope be designated the type specimen for Homo sapiens. Cope was a famous paleontologist who had willed his bones to science in the hope that he would be designated the type specimen (again, technically the lectotype) of Homo sapiens. And it certainly seems that Cope himself had wanted that: a possibly apocryphal story goes that a visitor to Cope’s laboratory shortly after the latter’s death in 1897 found his long-time technician weeping in front of a boiling preparation vat as Cope’s head periodically bobbed to the surface.

For anyone who cared, having two pretenders to the status of type specimen for Homo sapiens created a legally awkward situation that demanded resolution. Fortunately, the ICZN was up to the job. One rule states that a neotype can be assigned to a specimen if the lectotype is lost; and this might have given Cope’s bones a fighting chance as a neotype. But while Linnaeus is not exactly available, he is not exactly lost, for we know where he is. And his claim is reinforced by a couple of other provisions of the ICZN. One of them is that article 74.1 of the code happens to require that any lectotype must be among the specimens examined by the person who named the species. Linnaeus was long dead when Cope was born in 1840, so Cope could not have been “examined” by the namer. Also, the key article (74.1.1) clearly states the principle of priority under which validly proposed earlier names trump names put forward later. So unless Linnaeus’s name for our species is somehow deemed invalid—which, even in an uncertain world, is not going to happen—Cope’s claim doesn’t stand a chance. Meanwhile, Homo sapiens still lacks a usable type specimen.

This digression nicely illustrates the fact that, while systematists will always legitimately disagree, nomenclature is underpinned by an objective set of rules that we must apply in retrospect to Linnaeus’s four nonimaginary variants of Homo sapiens. The Swedish savant used highly typological reasoning to come up with what we would have to call subspecies, although he probably thought of them as races. And typologically, Linnaeus clearly felt justified in designating his four “real” subspecies based on geography and the skin color of the people involved. He also added some behavioral descriptions that he felt were diagnostic. These were pretty much in line with common European suppositions of the eighteenth century, and paramount were the ways in which the various groups controlled their behaviors. The “rufus” (red) Homo sapiens americanus from the New World used custom to govern its behavior; the “albus” (white) europaeus was governed by laws; the “luridus” (sallow) asiaticus from Asia was opinionated; and the “niger” (black) afer from Africa was impulsive. This blatantly racist and typological view of humans was hardly unusual for its time, and it remains significant as one of the first attempts to systematize the differences between human geographic groups.

In technical terms, Linnaeus’s trinomina stood until 1825, when Jean-Baptiste Bory de St. Vincent decided to elevate the subspecies names to species level and to add a raft of other regional populations to the genus Homo as separate species. But—to cut short a very long story that you can read about in our Race? Debunking a Scientific Myth—later experts have synonymized all of these with Homo sapiens, so that today no living Homo sapiens subspecies are recognized. Even Homo sapiens sapiens is entirely superfluous, since we have nothing to distinguish it from. Today, then, while we must revere Linnaeus for his achievements as a taxonomist, we must also admit that his splitting of the species Homo sapiens into geographic subspecies was the start of a hugely problematic—and hard to reverse or stamp out—trend toward the formal classification of human individuals and populations. Our colleague Jon Marks has suggested that this desire for racial classification has had an even greater impact on our modern life than the binominal system itself.

The Economist recently estimated the global pool of supernormal profits (profits in excess of an assumed 12 percent hurdle rate) at $660 billion. This is essentially a transatlantic phenomena; an astounding 98 percent of these excess profits accrue to a handful of American and European firms. In the United States, a third is cornered by firms with legal and regulatory moats such as healthcare, military contracting, and other non-tradables. But two-thirds go to firms in industries with no such artificial barriers to entry (tech and other tradables) and that are more or less fully exposed to competitors from across the world. Why aren’t these rents bid away by the entry of hungrier rivals?

A major part of the answer is network externalities that turn the logic of market competition over its head. To wit, if everyone is on Facebook you have to be on Facebook too if you want to connect with others. This applies to everyone on Facebook. Another social network trying to break through Facebook’s moat, even one unambiguously superior to the surveillance platform, faces the insurmountable hurdle of getting everyone to coordinate their move. The result is a natural monopoly.

Yet, that can’t be the whole story. Most firms that exhibit persistent rents do not operate in an industry with decisive network externalities or regulatory moats. Amazon is an attractive platform for people trying to buy and sell stuff. And this part of Amazon’s business indeed exhibits network externalities. But that business is not a cash cow for Amazon; third-party sales account for only 17 percent of Amazon’s top line. Instead, Amazon makes most of its money by selling stuff directly to customers. Jeff Bezos is now the richest man in the world with an estimated net worth in excess of a hundred billion dollars.

The secret of Amazon’s success is that it is the world leader in supply chain management. Three billion products are shipped worldwide by Amazon every month. The Balance calls the firm’s operations the most efficient in the world:

Shoshana Zuboff, ‘one of our most prescient and profound thinkers on the rise of the digital,’ Andrew Keen informs us on the blurb, has rendered ‘a book of immense ambition and erudition’. Naomi Klein thinks ‘everyone should read this book as an act of digital self-defense’. Joseph Turow concludes that from now on, ‘all serious writing on the internet and society will have to take into account The Age of Surveillance Capitalism.’  Robert B. Reich’s blurb calls it ‘a masterpiece of rare conceptual daring’, wherein she ‘demonstrates the unprecedented challenges to human autonomy, social solidarity, and democracy perpetrated by this rogue capitalism.’

It is hard not to concur with them all. Zuboff has identified a distinct regime of accumulation that emerged from the wreckage of the Internet stock bubble when first Google and then Facebook articulated the lucrative commodification of mass surveillance data. In doing so they have enclosed a vast new frontier without permission.  Zuboff goes so as to declare that surveillance data (“behavioral surplus” is her neologism)  is a fourth fictional commodity after Polanyi’s Land, Labor and Money (p. 100):

Today’s owners of surveillance capital have declared a fourth fictional commodity expropriated from the experiential realities of human beings whose bodies, thoughts, and feelings are as virgin and blameless as nature’s once-plentiful meadows and forests before they fell to the market dynamic. In this new logic, human experience is subjugated to surveillance capitalism’s market mechanisms and reborn as “behavior.” These behaviors are rendered into data, ready to take their place in a numberless queue that feeds the machines for fabrication into predictions and eventual exchange in the new behavioral futures markets.… In this future we are exiles from our own behavior, denied access to or control over knowledge derived from its dispossession by others for others. Knowledge, authority, and power rest with surveillance capital, for which we are merely “human natural resources.”

Her narrative is at its most powerful when later in the book she ties Facebook’s addiction machine to Natasha Dow Schüll’s masterpiece on machine gambling in Las Vegas. After documenting the scale of addiction to the social network and the corporation’s efforts towards that instrumental goal, she notes chillingly (p. 466),

All those outlays of genius and money are devoted to this one goal of keeping users, especially young users, plastered to the social mirror like bugs on the windshield.

But Zuboff’s account suffers from two flaws. The first, non-fatal flaw is that she drops the ball on a number of important threads that ought to have been more vigorously pursued. We get bare glimpses of the ecosystem of surveillance capitalism and great power intelligence agencies on the inside of the one-way mirror. Ditto the Obama-Google revolving door. The second, fatal flaw is that she drops the ball on the very mechanism of accumulation she identifies with such clarity. Zuboff brackets such potentially fruitful interrogation into the technical discourse of the high-priests of machine learning—presumably beyond her and the reader’s comprehension. That is very unfortunate. So on the prospects of solving the problem of causal inference from mass surveillance data, we don’t hear a word. She takes the 52 data scientists she interviewed at their word when they echo Silicon Valley’s visions of absolute power.

The fundamental problem with Zuboff’s account then is that she actually buys Skinner’s presumption that sufficiently high resolution surveillance data can and will allow those on the right side of the panopticon to predict the behavior of those on the wrong side with near-perfect fidelity. In reality, whether the problem, and it is a scientific problem, is even solvable in principle is not known. Indeed, there are good reasons to believe that it is not.

In fact, the same kind of problem plagues machine learning, molecular anthropology, and brain science (see Marsh’s sober and excellent take). In all three fields, predicting human behavior from fine-scale Big Data (respectively from mass surveillance, DNA and brain activity) has defeated all attempts. What you have in all cases is a massive torrent of data with very little by way of maps.

Imagine that you could identify the life histories of all fish in the world ocean right down to the last detail and suppose that you have access to free and unbounded computational resources (an impossibility). Even so, there is no reason to believe that there exist algorithms that could be discovered (whether by machines or human geniuses) that will allow you to predict with certainty which fish in a given school the shark will catch. All that may be (and is known to be) achievable is probabilistic statements (say relating the probability of being caught to the position of the fish in the school). Arbitrary amount of data cannot solve the problem of coming up with a theory of fish behavior, even if such a theory were to exist—not that we have any good reason to believe that it does.

As David Deutsch has argued, all scientific problems require specific solutions whose existence cannot simply be assumed. Even assuming that the problems are solvable in principle, we cannot know how many years, decades, centuries, or millennia it will take to solve them. Radical uncertainty is the calling card of original scientific research. What all this means is that it is entirely possible, nay likely, that almost all hopes for predicting human behavior with high fidelity projected onto surveillance data, DNA, brain scans, and so on, may very well be dashed. Simply put, Google cannot predict the next sentence I will write — even if it trains its algorithms on all texts ever authored by the human race including myself — because it does not have a theory of my mind, something that cannot be reverse engineered from surveillance data.

The underlying reason for sober pessimism on all such questions is well understood. Put simply, the reductionist paradigm — and all such hopes are based on it — is known to be of limited value in light of the overwhelming evidence for emergent complexity. Higher order open systems, where all the interesting causal vectors are located, defeat the search for a general theory in detail. That is, each higher level of analysis has its own properties and internal logic that simply cannot be identified at lower levels of analysis.  And high fidelity prediction requires identifying the correct causal structure at all relevant levels of analysis simultaneously. But we have no reason to believe that we can, or ever will be able to, identify all relevant levels of analysis; much less correctly identify their causal structure.

So, no, Google is not on the cusp of world domination. Google insiders’ ‘physics of clicks’ is best seen as intellectual bluster, not the harbinger of an automated society as Zuboff would have us believe.

Zuboff poses three fundamental questions throughout the book — Who knows? Who decides? Who decides who decides?

Who knows? Those on the right side of the one-way mirror have a clear asymmetric information advantage over those on the wrong side. But how much do they actually “know”? What they have is an ocean’s worth of surveillance data that they can probe with pattern recognition algorithms. This gives them some handle on broad patterns (ie they can build models and test them against the data) and they can examine the life world of “a person of interest” with a fine-toothed comb. But they can neither predict the future evolution of the broad patterns with much more certainty beyond what could already be done in the world of small data, nor can they predict the future behavior of the person of interest with greater confidence than a serious intelligence analyst without access to full-spectrum surveillance. So yes, we must recognize these new technologies of power and their potential for abuse. But let’s not pretend that they are all powerful. They are not: Data science is just fancy regression and ubiqutious surveillance is just fancy wire-tapping.

Who decides? Surveillance capitalists have laid claim to ownership of our vast digitally visible behavior. Just because they discovered the continent does not mean that they get to keep it. I second Zuboff’s call to make a sustained political effort to reverse the enclosure. What has been illegitimately dispossessed can indeed be repossessed. What we need is big push towards public oversight; one that goes beyond questions of monopoly and privacy and aims to essentially reverse the commodification of our behavior and ban automatic compliance (Zuboff’s “means of behavior modification”).

Who decides who decides? We can again agree with Zuboff that the discourse of “inevitability” is a transparent attempt at depoliticization. As she shows, other paths were indeed imagined; as when Google unveiled plans for Google Home where the householder was to have complete control over access to the data generated by connected devices. A different road was taken where your devices (“the Internet of Things”) are designed to generate surveillance data to be sold to those with a pecuniary interest in your future behavior. As her account shows, these decisions were taken unilaterally by technology firms under financial market pressure. They did not ask legal if it was compliant with the law and regulation. They just went ahead and did it. They got away with it, she argues, by moving at breakneck speed and under the radar; helped by the erosion of oversight institutions (especially the courts) under the neoliberal onslaught; and helped along too by the convergence of interests with the intelligence community after 9/11. It is now clear that this is politico-ethically unacceptable. Again, the onus is on us to mobilize politically to take back control of our digital footprint.

Zuboff mentions Andrew Tutt’s suggestion for an “FDA for algorithms” in passing. That is perhaps the single most constructive solution on offer and it is surprising to find her never returning to it. What is required is the incubation of a situated community of skilled software engineers and data scientists who can exercise effective oversight over their commercial counterparts in the public interest. An “FDA for algorithms” is precisely the right model for such an undertaking.

The Age of Surveillance Capitalism is compulsory reading. Zuboff has contributed greatly to our understanding of this rogue capitalism above all in identifying this specific regime of accumulation. So it is almost physically painful to discover that she has herself fallen for Silicon Valley’s Kool-Aid.