Yesterday, the Supreme Court ruled that credit card provider American Express’ long-standing policy of including anti-“steering” clauses in its contracts with merchants was not anti-competitive.

Steering is the practice whereby merchants discourage customers from paying with comparably high-cost cards like Amex and to use Visa or Mastercard instead. Importantly, anti-steering agreements do not limit merchants’ ability to favor debit cards or cash in their dealings with customers. The majority opinion, drafted by Justice Clarence Thomas, argues that there was no evidence of consumer harm in the form of higher prices or reduced output as a result of Amex’ anti-steering requirements.

The Court notes that the credit-card market is a market for transaction services and is two-sided, involving two distinct and interdependent sets of customers: merchants (who sell goods and pay fees to credit-card providers) and buyers (who use credit cards to pay for merchants’ goods).

Two-sided markets require a somewhat more sophisticated analysis in antitrust cases because of the interdependence of each side. For example, in a one-sided market, increasing merchant fees might be regarded as evidence of an abuse of monopoly power. In a two-sided market, such a fee increase can in fact be pro-competitive if it leads credit-card providers to offer better service to the other side of the market (buyers) thus increasing their purchases from merchants.

The peculiarities of two-sided markets have been noted by a long line of economists, including the 2014 Nobel Prize laureate Jean Tirole. His analysis, among others, was cited by Justice Thomas in his opinion.

Indeed, the evidence suggests that competition in the credit-card market is thriving. Take-up of cards has boomed since the 1970s. There is an endless variety of offerings among the various providers, with different (or no) annual charges, a diverse menu of rewards programs, and of course a range of interest rates on outstanding balances. In March, Amex in fact announced that it would cut its merchant fees to better compete with lower-cost providers. Since 2004, Amex fees have dropped by 10 percent.

However, even allowing that anti-steering agreements are not anti-competitive, could it be true nonetheless that anti-steering agreements hurt particular groups of consumers? Brookings’ Aaron Klein thinks so. In a piece published in the wake of yesterday’s ruling, he argues that “[the U.S.] payment system […] rewards the wealthy while penalizing the poor” and “functions as a hidden method of increasing income inequality.”

Klein’s argument is that anti-steering agreements lead merchants to increase prices. But because only Amex cardholders secure the countervailing benefits – in the form of rewards programs and an increased ability to use their cards in daily purchases – and because Amex cardholder incomes tend to be higher than average, there is a regressive impact on lower-income households. They get the higher prices induced by Amex fees but cannot share in the benefits.

Leave aside for a moment that anti-steering agreements, as Justice Thomas observed, only apply to credit cards and not to other forms of payment, which means that merchants can in fact price-discriminate by setting minimum purchase thresholds and surcharges for using credit cards (as many do). Even ignoring this possibility, which Klein leaves unacknowledged, his claim of regressivity is not as straightforward as might at first be apparent, for several reasons.

1. Merchants’ decision to sign anti-steering agreements is voluntary.

Merchants do not forcibly sign anti-steering agreements with Amex. They only do so if they believe it is in their benefit to accept Amex cardholders, which is a function, first, of the fees charged by Amex compared to other card networks, and, second, of the share of purchases paid for with Amex cards which would not happen if the merchants refused to take Amex.

The emphasis is important because many cardholders multi-home, meaning that they hold Visa or Mastercard as well as Amex for instances in which one might not be accepted. My dry cleaner does not take Amex but that has not yet deprived her of my custom.

Thus, merchants will only agree to the anti-steering conditions if the income earned from Amex purchases which could not otherwise be earned exceeds the cost of higher Amex fees. Many merchants will find accepting Amex worthwhile, but some will not. Indeed, there are still many outlets where all credit cards but Amex (and possibly Discover) are accepted.

In a competitive market with many providers and low switching and search costs (all of which broadly characterize most American retail markets), merchants will ask for the same price, adjusted for convenience factors such as the acceptance of Amex cards. Those who buy at Amex-positive outlets will pay a premium for it, while those who forgo the opportunity to buy with Amex can opt for other providers, who – other things being equal – will offer lower prices.

2. Amex cardholders and non-cardholders are different people, buying different things.

As Klein’s piece notes, lower-income households tend to have fewer credit card options, if any, than higher-income households. But to infer that the worse-off are therefore subsidizing the better-off because all pay the same price, but only cardholders get the rewards, is one simplification too many.

Households with different incomes differ not only in their likelihood of holding an Amex card, but also in the kinds of stores they patronize and the kinds of products they buy. This means that merchants have some means to make Amex cardholders internalize the price externality they would otherwise impose on non-cardholders, by passing on the Amex fees mainly, or exclusively, to those products that Amex cardholders buy.

Any pass-through will be crude because there is always overlap between the products bought by each of the two groups, but the homogeneous price increase for all customers posited by Klein is unlikely to reflect merchants’ reaction in practice.

3. There may be progressive redistribution among Amex cardholders.

Klein objects to the supposedly regressive impact of anti-steering agreements on those who do not hold Amex cards. But the existence of Amex rewards programs made possible by the anti-steering rules may be progressive, that is, it may redistribute benefits from higher-income to lower-income cardholders.

To see how this can be the case, consider that Amex cards tend to come with a “welcome bonus” involving, usually, a disproportionate amount of rewards (air miles, or gas points, or something else) for the first $1,000 spent on the card. These rewards are presumably as costly for Amex to give as any other rewards, so that the welcome gift must be subsidized from Amex’ ongoing business. The more households spend on their card above and beyond the first $1,000, the more they are subsidizing other cardholders. Furthermore, because card expenditure is broadly correlated with income, it is plausible that higher-income cardholders subsidize lower-income cardholders, for whom the first $1,000 are a bigger share of lifetime card expenditure.

This redistribution would offset the regressivity of steering restrictions on other buyers, to the extent there is any.

4. Innovation depends on a share of initial customers’ paying higher prices.

Innovations are always costlier at first. F.A. Hayek noted in The Constitution of Liberty that, without intending it, the rich who can afford new innovations help to make them progressively more accessible by encouraging investment and competition in the provision of the innovative good or service. From the personal computer to the smartphone to the transatlantic passenger flight, modern life is rich with examples of such gradual expansion in people’s access to new goods and services.

Credit cards are no exception. Justice Thomas noted that, since the 1950s when credit cards were first introduced, merchant fees have dropped by more than half. Because of the network effects characteristic of two-sided markets, every decrease in fees has likely disproportionately increased the amount of merchants accepting cards and the number of cardholders. The result is higher welfare for all and more widely shared affluence.

But this virtuous cycle depends on the ability of competitors to offer different price and quality options to prospective customers. Amex’ business model of higher merchant fees in exchange for more generous rewards programs is one form that such competition can take.

Klein concludes his piece by noting that financial technology can resolve the “inefficiencies of the current payment system [which] cry out for new financial technological solutions.” In this he is doubtless correct, if the history of financial innovation is anything to go by. Yet, to blame American Express for the remaining imperfections in the payment system would not just be wrong — it would be counterproductive.

[Cross-posted from Alt‑M.org]

Should cryptocurrencies be regulated like securities? Financial regulators have been pondering this question for some time. In a Briefing Paper published today by the Cato Institute’s Center for Monetary and Financial Alternatives, I suggest that securities regulation would only seem warranted in certain clearly circumscribed cases. For the most part, cryptocurrencies should be treated like commodities.

It has been nearly a decade since “Satoshi Nakamoto” laid the intellectual foundations for Bitcoin, the first cryptocurrency platform. Since then, more than 1,600 peer-to-peer networks have emerged to disrupt established intermediaries. Cryptocurrencies, even in the comparably bearish first half of 2018, have an aggregate market capitalization of nearly $300 billion.

While policymakers’ attention has gradually turned to designing an appropriate regulatory framework for this emerging technology, policy uncertainty persists. On one hand, some policymakers recognize the potential for cryptocurrencies to increase competition, reduce transaction costs and improve capital formation opportunities for firms. On the other, statements from regulators at the SEC and CFTC, the two agencies most closely monitoring the development of cryptocurrencies, have been unclear, equivocal, and sometimes outright contradictory.

In April, former CFTC chairman Gary Gensler suggested that ether, the cryptocurrency of the Ethereum network, should be treated like a security. This designation would have forced onerous new registration requirements on platforms that hold ether in custody and for trading, while making access to Ethereum by retail buyers more difficult. Furthermore, because the Ethereum platform provides the infrastructure for many other cryptocurrencies, such a move would have compromised the viability of large parts of the cryptocurrency market.

Fortunately, these risks have receded into the background since William Hinman, Director of the SEC’s Division of Corporation Finance, argued in a recent speech that ether, in its present form, wouldn’t qualify as a security because the Ethereum platform is heavily decentralized. The SEC’s classic test for a security defines it as (1) an investment of money (2) in a common enterprise (3) with the expectation of profits (4) from the efforts of others. In Hinman’s opinion, developments since the launch of Ethereum in 2014 mean that the platform presently fails to meet criteria (2) and (4).

The next step is to give Hinman’s welcome pronouncement regulatory heft. In the Briefing Paper released today, I propose that the SEC and CFTC formally establish a distinction between functional cryptocurrencies, such as Bitcoin and Ethereum, and promises of cryptocurrencies to be delivered in the future. Cryptocurrencies in the first category do not meet the criteria for a security and should be regulated as commodities. Those in the second category may be securities, depending on the individual circumstances of each issue.

The launch of cryptocurrencies is often preceded by what is, somewhat misleadingly, called an initial coin offering (ICO). An ICO involves the exchange of money today for the delivery of units of cryptocurrency in the future, where the funds are used to build a new platform. ICOs are a way for startups to raise capital, so in some circumstances they may tick the four boxes in the SEC’s security test. In particular, when buyers in an ICO are able to trade their holdings before the launch of the application, the contracts could constitute securities. In other cases, however, ICO agreements may simply be advance purchases of a good or service and not tradable before the platform goes live. Those agreements more closely resemble forward contracts and should be regulated like them.

In the paper, I propose just such a two-tier regulatory structure for ICOs, recognizing that some of them may fall under the securities laws, but that this will be determined by the circumstances of each case.

Apart from being consistent with Director Hinman’s position, the suggested approach balances consumer protection and the duties of financial regulators with an open environment for cryptocurrency innovation. It recognizes that most of the fraud about which the SEC has expressed concern happens at the ICO stage, so buyers might benefit from increased disclosures then. But it also takes account of the fact that excessive regulation of functional cryptocurrencies would stifle the market and throw a spanner in its further development, with few countervailing benefits in the form of market stability or consumer protection.

It is time for policy to catch up to the exciting development of cryptocurrency markets. But catching up shouldn’t mean smothering the technology with regulation, nor crudely applying the securities laws to all cryptocurrencies. The reality of this emerging market favors a more judicious approach.

[Cross-posted from Alt‑M.org]

On Twitter last week Stephen Williamson wrote that he was “puzzled by the infatuation with NGDP targeting. We have good reasons to care about the path for the price level and the path for real GDP. Idea seems to be that if you smooth Py that you get optimal paths for P and y. That’s hardly obvious, and doesn’t fall out of any serious theory I’m aware of.”

I’m not exactly sure what Stephen means by a “serious theory.” But if he means coherent and thoughtful theoretical arguments by well-respected (and presumably “serious”) economists, then there are all sorts of “serious theories” out there to which he might refer, with roots tracing back to the heyday of classical economics. Indeed, until the advent of the Keynesian revolution, a stable nominal GDP ideal, or something close to it, was at least as popular among highly-regarded economists as that of a stable output price level, its chief rival.

If one of today’s outstanding monetary economists isn’t familiar with these writings, it’s a good bet that many other economists don’t know about them either. So I thought it worthwhile to list here some important theoretical works favoring nominal GDP targeting over inflation or price-level targeting, in chronological order, with brief remarks concerning each, and links to those that can be read online. Those interested in a more complete survey of the relevant literature, albeit one written not quite a quarter-century ago, are encouraged to have a look at my 1995 History of Political Economy article, “The ‘Productivity Norm’ versus Zero Inflation in the History of Economic Thought.”

Astute readers will note that the works I list are ones that compare the fundamental welfare effects of stable aggregate spending to those of other monetary policy objectives. They do more, in other words, than merely suggest that nominal GDP may be a useful intermediate policy target than, say, some monetary aggregate. The articles also reject absolute stability of the price level, real output, or both as “given” policy ideals. In contrast, some other writings favoring nominal GDP targeting (e.g. McCallum 1987, Hall and Mankiw 1994, or Frankel and Chinn 1995) do so because it does a reasonably good job at stabilizing P or y or some weighted average of the two. These writings seem to me to miss the crucial point that output and price level fluctuations are themselves sometimes optimal, and that NGDP targeting is in turn desirable precisely because it allows such optimal fluctuations in y and P to occur. Those critics of NGDP targeting (e.g., George Kahn and Andrew Levin) who fault it for failing to stabilize P and y, as well as some alternative policies, likewise miss the mark. To all of these writers I especially commend the works listed here.

Finally, I leave out works by Scott Sumner, David Beckworth, and other well-known Market Monetarists, because part of my aim is to show my fellow monetary economists that Market Monetarism is but one particular, recent manifestation of a perspective that has had many distinguished adherents, representing numerous different schools of thought, throughout the long history of our discipline.

***

1837: Samuel Bailey, Money and Its Vicissitudes in Value

A splendid, early tour de force. Bailey may have been the first economist to distinguish between changes in the value of money (and, hence, the general level of prices) “originating on the side of money” and ones “originating on the side of goods” — where by the latter he means changes reflecting innovations to the “facility of supplying” goods, including “the discovery of shorter and more economical processes in the arts, and invention of machinery, [and] the abolition of monopolies and taxes.” Today’s economies make essentially the same distinction when they trace price-level changes to either “aggregate demand” or “aggregate supply” innovations. Using this distinction, Bailey goes on to show that, unlike price level changes originating on the “money” side, those originating on the “goods” side neither distort nominal (“pecuniary”) contracts nor (in the case of downward price movements) discourage industry as a whole.

1918–1933: Debate in the Ekonomisk Tidskrift

In a series of articles in the Ekonomisk Tidskrift, Stockholm School economist David Davidson took his compatriot, the great Knut Wicksell, to task for identifying a neutral (hence ideal) monetary policy with one that achieved a perfectly stable general price level. Davidson instead insisted that a neutral policy would allow the price level to vary with aggregate supply (and, particularly, productivity) innovations. Among other prominent Stockholm-School economists, Eric Lindahl and Gunnar Myrdal (see below), sided with Davidson in this debate, while Gustave Cassel sided with Wicksell. Regrettably, Davidson’s works haven’t been translated into English. But those who don’t read Swedish may consult excellent reviews of the debate by Brinley Thomas and Claes-Henric Siven. According to Thomas, “There can be little doubt that the honours in this contest went to Davidson.” A recent paper by Gunner Örn, also (alas) in Swedish, points out the practical equivalence between Davidson’s norm and recent proposals for targeting nominal GDP. “These two Swedish economists,” he writes (referring to Davidson and Lindahl), “were very early — perhaps even the world’s very first — advocates of what in today’s economics debate is called ‘nominal GDP targeting’.”

1930: Ralph G. Hawtrey, “Money and Index-Numbers”

If this essay were all Hawtrey ever wrote on monetary theory, Scott Sumner might still have had reason for wanting his chair at Mercatus to be called “The Ralph G. Hawtrey Chair of Monetary Policy.” For no-one has ever made the case for preferring a stable level of nominal income (or what Hawtrey refers to as “consumers’ income and outlay”) to a stable price level more clearly and painstakingly than Hawtrey makes it here. The only disturbances to prices that monetary policy should seek to avoid, Hawtrey concludes, are those “affecting the amount of the consumers’ income and outlay otherwise than in proportion to the factors of production.” Spot on, Ralphie boy!

1935: F.A. Hayek, Prices and Production, 2nd. ed.

Let’s be clear: my concern here is with writers who made compelling theoretical arguments for NGDP targeting, or something very close to it, regardless of whether they believed the theory could be put into practice. By this criterion, Hayek must be judged one of the more important historical precursors of Market Monetarism. Although Hayek had been critical of arguments and schemes for stabilizing the general price level for some years before he wrote Prices and Production, it was only in that work, and particularly in it’s 2nd edition, and in several subsequent writings, that he acknowledged the desirability of such money-stock changes as would serve to insulate the price level from velocity (but not real output) shocks. For further details, and a comparison of Hayek’s views on this issue with Keynes’s, see my 1999 HOPE article. Regular Alt‑M contributor Larry White has written a good, sympathetic review of Hayek’s monetary thought. Rather less sympathetic is a blog post by “Lord Keynes” pointing out differences between Hayek’s stable MV ideal and his practical policy recommendations during the Great Depression.

1937: Allen G. B. Fisher, “Does an Increase in Volume of Production Call for a Corresponding Increase in Volume of Money?”

Unlike his better-known American namesake, this notable New Zealand economist denied that a stable price level was always desirable. Instead he insisted that, “Apart from increases in population and from changes in the desire of individuals to hold money, economic development which takes the form of increased production per head…does not require any increase in money supply.” In other words, he favored an NGDP target with the target growth rate of NGDP set equal to the population growth rate. It is not falling prices per se but generally falling money incomes, Fisher argued, that “have a depressing effect upon business enterprise” because they disappoint “expectations of normal profit.” A decline in prices reflecting increased productive efficiency is, on the other hand, not only not harmful, but desirable, because attempts to prevent them is likely temporarily to distort profit signals, delaying desirable transfers of scarce resources among different efficient industries. Highly recommended.

1939: Gunnar Myrdal, Monetary Equilibrium

Many people know that Hayek and Myrdal shared the 1974 “Nobel” prize “for their pioneering work in the theory of money and economic fluctuations” (among other things). But relatively few have read Myrdal’s main work on monetary theory, and even fewer realize that it was originally published, in German, for a 1933 volume that Hayek edited. Least well known of all is the fact that Hayek didn’t think much of Myrdal’s analysis at the time, and that he accepted it only as a last-minute substitute for one that Eric Lindahl was supposed to prepare. For that and other (mainly ideological) reasons there was no shortage of bad blood between the two men when they crossed paths again in Stockholm. Still, nothing could excuse Myrdal’s boorish behavior both during and after the event.

Having gotten all that off my chest, the fact remains that, whatever its flaws, Myrdal’s essay is well worth reading. Among other points, Myrdal argues — convincingly, I think — that “If one desires the greatest possible diminution of the business cycle, but at the same time wants a guarantee against too great, and especially unidirectional, price movements, which naturally affect distribution most severely, then one must try to stabilize an index of those prices which are sticky in themselves. This would often lead in practice to a stabilization of wages.” In this regard Myrdal’s argument anticipates some more recent New Keynesian arguments favoring a stable NGDP growth trend over a stable price level or inflation rate.

1963: Dennis Robertson. A Memorandum Submitted to the Canadian Royal Commission on Banking and Finance

For better or worse (the last, if you ask me), Keynes’s General Theory made a clean sweep of the field of monetary economics, brushing aside competing works, including the insightful contributions of Keynes’ s Cambridge colleague. Robertson’s popularity thereafter was largely sustained by his splendid little handbook on Money, first published in 1922 and revised and reprinted many times thereafter. That’s a shame, because Robertson’s less popular works on the same subject are full of brilliant insights, including some exposing fundamental errors in Keynes’s, if not in Keynesian, monetary theory.

Robertson’s “Memorandum,” which he completed in 1962 but which was not published until a month after his death in 1963, was his last important work on monetary economics. In it he asks “how a Monetary Authority should behave in a country which was isolated from the rest of the world, and in which it was desired so far as possible to leave the pace of capital formation to be determined by the unfettered interplay of the decisions of private enterprisers and savers.” His answer is that the authority’s aim should be to stabilize, not the general price level, but the economy’s “money flow,” meaning “the flow of monetary demand for final output.” Doing so, he argues, would better “enable the participants in the growth process — enterprisers, savers and hired workers — to realize their intentions with a minimum of friction and of distortion of the true significance of the monetary contracts which they are making with one another.”

1983: Charles Bean, Targeting Nominal Income: An Appraisal

The monetarist counterrevolution, with its emphasis on targeting monetary aggregates, ultimately helped to inspire a new round of arguments for targeting nominal income, with contributions by James Meade (in his 1977 Nobel address), James Tobin, and Sam Brittan. Bean’s assessment of them is highly analytical, and in that respect at least especially “serious.” He concludes

that a policy of targeting nominal income is an optimal response to demand shocks and to productivity shocks if labour supply is inelastic. Even if labour supply is elastic nominal income targets will still produce a better response to productivity shocks than monetary targets if the price elasticity of aggregate demand is less than unity. Growth rules are less attractive than targets for the level of nominal income.

Given the quality of this early performance, it’s a shame that Bean chose to follow it up more recently with a far less compelling speech pooh-poohing recent arguments for targeting nominal GDP as so much “old wine in a new bottle,” and otherwise attempting to suggest that, had it been pursued during the crisis, such targeting would not really have improved much on the U.K.‘s inflation targeting regime. Why “less compelling”? Here’s an example: Bean at one point argues that, because “the financial crisis has led to a fall, possibly temporarily, in the underlying rate of growth of supply,” a given or “fixed” nominal income target would have gone “hand-in-hand with a higher inflation rate, whereas a fixed inflation target would [have been] associated with lower nominal income growth.” Quite correct; and had Bean stopped there he’d surely have been compelled to conclude, as he had in 1983, that NGDP targeting would have been the better option. But instead of doing so he continues: “But this hardly provides an argument in favour of a nominal income growth target. Indeed, in this case one would surely want to set the target growth rate for nominal income lower to reflect the lower rate of growth of supply, though by how much might be hard to judge.”

Huh? What? Here we see the dangers lurking in any comparison of alternative monetary policy norms that assumes (as Bean’s does here) that deviations from some target inflation rate belong in the monetary authority’s “loss function.” The whole point of NGDP targeting is that it allows supply-side developments, and adverse ones especially, to be reflected in higher prices or (in the case of a drop in the growth rate of productivity) a higher inflation rate. A central bank that sets a target for NGDP growth only to revise it in response to supply innovations so as to maintain a constant inflation rate is targeting inflation, not NGDP.

1989: Michael Bradley and Dennis Jansen, “Understanding Nominal GNP Targeting”

Short, simple, and sweet: the framework is good old aggregate supply and demand, with diagrams and all. For a bonus, there’s a nice discussion of Ben McCallum’s NGDP Rule. Read this before you try to work your way through either Bean’s 1983 article or the New Keynesian works listed further on.

1992: William Niskanen, “Political Guidance on Monetary Policy”

Although he’s mainly known for his contributions to the theory of bureaucracy, and to public choice theory more generally, Bill Niskanen, Cato’s long-time chairman, was an all-round original thinker. Nor did he neglect monetary economics, where his thinking led him straight into the nominal-spending stability fold. Like Robert Gordon before him, Niskanen differed from today’s Market Monetarists mainly in preferring a nominal “final sales to domestic purchasers” target to a nominal GDP target. (On the pros and cons of these alternatives, see this excellent Bill Woolsey post.)

2005: Jinill Kim and Dale Henderson, “Inflation Targeting and Nominal-Income-Growth Targeting: When and Why Are They Suboptimal?”

Using “a model of a closed economy with optimizing firms and households, monopolistic competition in both product and labor markets, and one-period nominal contracts,” Kim and Henderson “derive optimal monetary stabilization rules and compare them to simple rules under both full and partial information.” Importantly, by “optimal rules” they mean, not simply ones that minimize an ad-hoc central bank “loss function,” but ones that “maximize the unconditional expected utility of the representative agent.” They find that, while none of the simple rules quite match up to the optimal rule, among them “nominal-income-growth targeting dominates inflation targeting for plausible parameter values.” They also find, not surprisingly, that “the more important are productivity shocks…the greater the advantage of nominal-income-growth targeting over inflation targeting.” There’s nothing new in these particular conclusions, which (as we’ve seen) many previous economists arrived at with relatively little algebra, or with no algebra at all. Still, formal models like this one may prove essential to winning over those NGDP targeting doubting Thomas’s who believe nothing until they see it drop out of a fully-articulated macro model.[1]

2010: Evan Koenig, “The Case for Nominal Income Targeting”

As Senior Vice President and Principal Policy Advisor at the Dallas Fed, Koenig is among a small number of Fed insiders who have added their voices to others drawing attention to the possible advantages of targeting nominal GDP. He starts this article by listing three “desiderata” of a monetary policy rule. These are (1) that the rule should avoid waste that might otherwise result from “sluggish” prices or wage rates and other “frictions”; (2) that it should maintain low and stable long-term inflation expectations; and (3) that it should promote financial stability. He then proceeds to explain how a nominal GDP rule might achieve them all.

Concerning price and wage rigidities, Koenig’s argument is reminiscent of Myrdal’s: “If it is primarily money wages that are sticky, not product prices,” he observes, “then it will be optimal for monetary policymakers to try to avoid surprise changes in the market-clearing money wage and let product prices move up or down as needed to clear the labor and product markets.” Furthermore, “In an economy in which the principal real disturbances are shocks to productivity,” and “the income and substitution effects that productivity shocks have on employment are offsetting — a reasonable rough approximation — then the optimal wage policy is equivalent to targeting the level of nominal spending.” Koenig shows as well that “a nominal-income target for monetary policy distributes risk more efficiently across debtors and creditors than does a price-level target.” His overall conclusion is that “A level target for nominal spending…offers enough potential advantages relative to a simple price-level target that it deserves careful study.”

2014: Kevin Sheedy, “Debt and Incomplete Financial Markets: A Case for Nominal GDP Targeting”

Sheedy’s very thorough and carefully-reasoned study is concerned exclusively with the task of identifying a monetary regime best suited to an economy that relies on fixed (“non-contingent”) nominal debt contracts. As such it takes up a theme first addressed in depth by Samuel Bailey 80 years earlier. Notwithstanding the passage of so much time, and Sheedy’s far more formal framework, his conclusion is essentially the same, namely, “that when debt contracts are written in terms of money, a monetary policy of nominal GDP targeting improves the functioning of financial markets.” Commenting on Sheedy’s work, James Bullard writes that it “has considerable potential to sharpen the ongoing debate on nominal GDP targeting, an idea that has not often had an explicit modern macroeconomic model behind it.”

2016: Julio Garin, Robert Lester, and Eric Sims, “On the Desirability of Nominal GDP Targeting”

It’s nice to be able to conclude my survey with a paper written by my former colleague, Julio Garin, and two co-authors. As I observed in reviewing the paper by Kim and Henderson, theirs is only one of a number of papers that use sophisticated macro models to conclude that nominal GDP targeting beats most other simple monetary rules. Garin et al. “compare nominal GDP targeting to inflation and output gap targeting as well as to a conventional Taylor rule…on the basis of welfare losses relative to a hypothetical equilibrium with flexible prices and wages.” In other words, like Kim and Henderson, they don’t simple ask how NGDP targeting stacks up when it comes to stabilizing P or y or both — which isn’t the same thing.

Their findings? Although output gap targeting is generally ideal, nominal GDP targeting performs almost as well, and better than a Taylor rule. NGDP targeting is also a lot better than inflation targeting. Furthermore, it can even outperform output gap targeting if the output gap is observed with noise. Finally, the advantages of NGDP targeting over rival strategies increase as supply shocks become more pronounced, and in situations where wages are stickier than prices. If all this is starting to sound familiar, that’s called robustness.

***

That last remark compels me to conclude by answering a question that’s bound to occur to many who have read this survey, to wit: if the theoretical case for nominal GDP targeting is so strong, why do so many monetary economists favor price-level or inflation targeting?

There are, I believe, two reasons. One is that these economists have tended to assume, implicitly or otherwise, a world in which (1) aggregate supply innovations, and innovations to total factor productivity in particular, are either non-existent or unimportant; and (2) factor prices are perfectly flexible or, at very least, more flexible than output prices. In such a world, nominal GDP targeting may not have any decisive advantage over price-level or inflation targeting, though it’s also unlikely to be any worse than either.

The other reason is that many monetary and macroeconomists treat fluctuations in the price level or inflation rate as bad per se, most commonly by treating them as arguments in central bankers’ “loss functions.” To say that this procedure lacks solid microfoundations is putting things mildly. A less temperate reply might be something like, “Who cares what central bankers like or don’t like? It’s peoples’ utility or welfare that matters.”

In short, if there’s a shortage of “serious” theorizing about what central banks ought to stabilize, it exists, not among proponents of nominal GDP targeting, but among those who continue to favor price-level or inflation targeting.

_____________________

[1] For all their intricacies, there’s at least one respect in which such fully-articulated macro models still fail to provide an adequate framework for assessing the advantages of inflation or price-level targeting relative to those of targeting nominal GDP. So far as I’m aware, no one as yet has constructed a version of such a model in which the responsiveness of actual nominal prices to shocks that alter those prices general equilibrium levels depends on the nature of the shocks in question. In contrast, numerous less formal writings allow, for example, that money prices are much more responsive to supply (unit cost) innovations than they are to changes in demand. This last fact tends to further strengthen the case in favor of a stable NGDP rule.

[Cross-posted from Alt‑M.org]

The Federal Reserve plans to review its inflation growth rate target and potentially select a new monetary policy target. Many Fed officials are in favor of the idea, including Chair Powell. And the latest FOMC minutes show that the Fed’s policy setting committee has discussed new targets.

This is good news, because inflation rate targeting suffers from serious shortcomings. With a growth rate target, a central bank writes off past errors. Instead of deliberately correcting those errors, it “lets bygones be bygones,” allowing its mistakes to permanently alter the value of its policy target. For example, the Fed has persistently undershot its 2% inflation target since adopting it in January 2012. Because it does not plan to atone for this prolonged period of undershooting the price level will forever remain below its original target path, at least so far as deliberate Fed actions are concerned.

Successful price level targeting requires, on the other hand, that the central bank make up for past mistakes. Monetary policy is successful if the price level returns to the trend line it was growing along before the undershooting occurred. This makes the future course of the price level easier to predict. Inflation growth rate targeting cannot match this degree of predictability because its policy errors permanently change the long run price level, making the future path of the price level more like a random walk.

Improved price level predictability is one of the reasons that several Fed officials have discussed the benefits of adopting a price level target. For example, John Williams, who will take over as President of the New York Fed and become Vice Chair of the FOMC next Monday, argues that a price level target would offer an increased level of predictability over inflation rate targeting by better indicating where prices will be 5, 10, even 30 years into the future. Williams also believes, rightly, that this predictability would make future Fed policy more transparent to the public.

Richard Clarida, Vice Chair-elect at the Federal Reserve Board, has also suggested implementing a price level target. He made the suggestion back in 2014 because he was concerned that the Fed lacked the ability to credibly communicate the future path of policy absent a level target. At the time, the Fed was following the Evans Rule, a promise to leave interest rates at the zero lower bound until either inflation was above 2.5% or the unemployment rate was below 6.5%. Clarida saw that these thresholds were not enough to anchor the public’s expectations going forward. He feared that so long as the public’s expectations remained insufficiently anchored, the Fed would continue to struggle to meet its policy target.

In his most recent speech, the Atlanta Fed’s new President, Rafael Bostic, also endorsed a price level target as an alternative to the Fed’s current inflation rate target. Like Williams and Clarida, Bostic believes that price level targeting would have performed just as well as inflation rate targeting throughout the Great Moderation, and that it would have been superior to it since the Fed adopted an explicit inflation target in 2012.

While Fed officials are right to believe that price level targeting can improve upon inflation rate targeting, they fail to consider the shortcomings of either alternative in the presence of supply shocks. A price level target (to refer only to the better of these two options) may be optimal in the absence of such shocks, but in their presence it makes monetary policy procyclical.

Consider the case of a negative supply shock. A sudden fall in the global production of oil would likely push up domestic gas prices, which would in turn raise the price level. Such a rise would be a signal to the Fed to tighten monetary policy. Yet, tighter monetary policy would provide no relief to the economy in such a circumstance. Tighter policy would put further downward pressure on an economy whose consumers already feel constrained by higher prices because of the oil shock. Only if the rise in gas prices was the result of excess aggregate demand, something likely caused by over-easing by the Fed, would tighter monetary policy be appropriate.

Positive supply shocks can likewise have procyclical consequences. Were the United States to see a (welcome) improvement in productivity the inflation rate would tend to fall. After a short period with the lower inflation rate the price level would still be rising but be below its target path. Under a price level target, the Fed would respond with easier monetary policy in an effort to raise the inflation rate and bring the price level back up to its path. But prices falling because products are made more efficiently is a gain for consumers, who ought to enjoy lower prices on those products. Trying to raise the overall price level in an effort to “combat” these productivity gains should hardly be part of a central bank’s policy and could risk overheating the economy.

In short, the tendency of a price level targeting central bank to respond to positive supply shocks in the same way as it responds to negative demand shocks, and to respond to adverse supply shocks in the same way as it responds to positive demand shocks, is a recipe for trouble. To their credit, both Bostic and Williams discuss implementing a “flexible” price level target, one that could be adjusted to changing economic circumstances. Such flexibility would allow the Fed, at least in theory, to avoid procyclical monetary policy during supply shocks by allowing the price level to change.

But flexible price level targeting is really just a more ad-hoc, and therefore less robust, version of a nominal GDP level target. Nominal GDP is the overall size of the economy uncorrected for inflation, so nominal GDP growth is essentially the sum of the real growth rate and the inflation rate. Under a nominal GDP level target the central bank would be stabilizing overall spending, thereby automatically and systematically doing what flexible price level targeting is supposed to accomplish with less risk of implementing procyclical monetary policy.

Reconsider the previous example when the inflation rate tends to fall during periods of improved productivity, except now the central bank has a nominal GDP level target. With the inflation rate falling the price level would fall below its previous trend, but that decline would not elicit any procyclical response from the central bank. Under a nominal GDP level target the central bank only responds velocity shocks. The central bank would adjust the money supply to offset velocity shocks, in an effort to stabilize overall spending and keep nominal GDP growing on its trend.

Because it focuses the central bank’s response function on one variable, changes in velocity, a nominal GDP level target is the best target for monetary policy. On the other hand, a price level target, and its advocates fail to fully account for this, obligates the central bank to react to changes in velocity and changes in aggregate supply. A nominal GDP level target offers the same degree of predictability as a price level target, but has the additional advantage of being robust to supply shocks, precisely because it allows the price level to change. Under a nominal GDP level target, the chances of the Fed being procyclical during a downturn and amplifying the contraction would be greatly reduced.

Price level targeting proponents are right to believe that it is superior to inflation rate targeting because it corrects the bygones problem, improving the Fed’s performance by making the price level more predictable. However, a price level target is the ideal only in a world without supply shocks. With supply shocks, a central bank with a price level target would too often act in a procyclical manner. A “flexible” price level target is certainly a better option than a strict price level target. But it would only be the best available option if it were so “flexible” that it amounted to nothing other than a nominal GDP level target.

Fed officials are making the right decision to rethink the current inflation rate target. For that review to be successful, they should thoroughly consider, but ultimately reject a price level target. Instead, they should listen to a growing number of economists discussing the best option: a nominal GDP level target.

[Cross-posted from Alt‑M.org]

As if central banks’ powers and balance sheets haven’t grown quite enough since the outbreak of the subprime crisis, we’ve been hearing more and more calls for them to expand their role in retail payments, by supplying digital money directly to the general public.

Some proposals would have central banks do this by letting ordinary citizens open central bank accounts, while others would have them design and market their own P2P “digital currency.” Either sort of central bank digital money would, the plans’ supporters claim, be just as convenient as today’s dollar-denominated private monies. But central bank digital money would also have the distinct advantage of being just as safe as paper money.

Earlier this week the FT’s Martin Sandbu jumped onto the central bank “ecash” bandwagon, in an article prompted by the recent disruption of Visa’s European payments network. That disruption, Sandbu wrote, supplied “one of the strongest considerations in favour of introducing official electronic money.”[1]

Sandbu’s argument is just one of many that have been offered for allowing central banks to supply ecash. But it’s representative of the rest in at least one crucial respect: like them, it may seem solid enough at first glance. But upon closer inspection, it turns out to be full of holes.

Central Banks and Computer Glitches

Absent a crisis of confidence, the most likely causes of a private payments system disruption are (1) hacking and (2) a software or hardware breakdown. It appears that a computer hardware failure was to blame for Visa’s European troubles, although hacking was suspected at first.

Payments systems operated by central banks are similarly dependent on computer hardware and software, and are for that reason also vulnerable to both hacking and equipment failures. That’s the first — and far from trivial — flaw in Sandbu’s argument. Within the last two years, for example, hackers have used malware to steal millions from the central banks of Russia and Bangladesh. In the latter case the money came straight out of the Bangladesh Bank’s account at the New York Fed. Had it not been for a stroke of good luck, the bank’s losses — $101 billion, about a third of which was eventually clawed back — would have been far greater. During the same period hackers also managed to plant a digital “bomb” into the software of the Saudi Arabian Monetary System. Back in 2014, a computer failure at the Bank of England held up thousands of payments, including many by persons trying to close on new homes. So much for the perfect safety of central bank digital money.

If all electronic payments systems are vulnerable to computer-related failures, is there any way to protect oneself against them? In fact there are at least two ways. One is to avoid putting all one’s payments eggs in one basket, by keeping multiple credit cards and bank accounts, and by subscribing to PayPal or other independent payment service providers. Of course, keeping funds at a central bank would be another way to diversify, were it allowed. But with so many private-market options out there, it’s absurd to suppose that people can’t protect themselves from payment system glitches unless central banks themselves enter the electronic cash business.

The other option is to keep some good old paper money on hand. Moreover, that’s the only option, apart from resort to barter, that would help in the case of a truly global electronic payment system breakdown, however that might happen. (A cosmic ray shower, perhaps.) But far from being an argument for having central banks enter the electronic payments fray, this far-fetched scenario is a good reason for having them to stick to supplying paper money.

A Flight to E‑Cash?

Besides claiming that central bank ecash would protect its holders from the risk of a payments-system breakdown, Sandbu suggests that it would “force a move towards higher reserve requirements for banks,” and perhaps even toward full-reserve banking. Allowing central banks to supply digital money to the general public could, in other words, lead spontaneously to the same outcome proponents of the Vollgeld initiative are hoping to achieve in Switzerland by means of next week’s referendum. This would happen, Sandbu says, because the public’s ready access to such cash would result in “a massive flight from deposits to safer official money.” To allow for this contingency, without having to resort to massive last-resort lending, central bankers would have to see to it “that banks hold enough reserves for the purpose up front.”

Most people would consider a policy change that’s capable of triggering massive bank runs a bad idea. But so far as Sandbu is concerned, increasing the likelihood of such runs is just a convenient way to put paid to fractional-reserve banking, of which he evidently disapproves. Like most critics of fractional-reserve banking, he doesn’t say who will supply the credit commercial banks can no longer offer once they convert to a full-reserve basis. Also like them he appears to appreciate neither the synergies between deposit taking and lending that account for their coexistence since the beginnings of banking nor the fact that, if fractional reserve banking systems sometimes appear fragile and unstable even when not threatened by direct competition from central banks, we often have other misguided bank regulations (including under-priced government guarantees) to thank for it.

But would the mere appearance of central bank ecash really provoke “a massive flight from [private bank] deposits”? It’s true that, so long as they aren’t promising to peg their currencies to some other national currency, central banks can’t default: to break a promise, one has to make one in the first place. But given the widespread presence of deposit insurance, and the fact that certain banks are considered Too Big To Fail, most readily-transferable commercial bank deposits (the sort for which ecash is a close substitute) are either explicitly or implicitly insured. Of approximately $12 trillion in U.S. demand deposits, for example, just over $7 trillion are insured, while much of the remainder consists of deposits held at the very largest U.S. banks.

It follows that, if there’s to be a massive switch from from commercial bank deposits to central bank ecash, it will have to be inspired, not merely by that alternative’s safety, but by its other features, including its convenience and interest return.

Central Banks Make Poor Competitors

Might central bank ecash dominate privately-supplied alternatives along these other dimensions? It might, but only if central banks cheat.

Let’s start with interest. Commercial banks’ main business consists of attracting deposits and figuring out how to invest them profitably. Competition compels them to seek high risk-adjusted returns (or, if they’re Too Big to Fail, to seek high returns regardless of risk), and to share those returns, less their overhead and operating expenses, with their depositors. Central banks, in contrast, are not supposed to be looking out for high returns. Instead, their assets typically consist of relatively safe and low-yielding securities, high-grade commercial paper, foreign exchange, and gold. To the extent that central banks extend credit, they extend it (with occasional, and often controversial, exceptions) to financial firms only, not to earn a profit, but to secure financial stability. It’s owing in part to this crucial difference between central and commercial banks that any public substitution of central bank money for commercial bank money is likely to result in a decline in total lending.

Most monetary policy experts would not want to change these limitations on central banks’ ability to profit by their investments. Nor do I suppose that Mr. Sandbu is an exception. After all, to the extent that central bank portfolios resemble those of ordinary commercial banks, they cease to be particularly safe institutions; and even if holders of their liabilities are not themselves directly exposed to the risks they take, taxpayers are. Allowing central banks to emulate commercial banks, not only by being able to supply digital money to the general public, but by taking on similar risks, would defeat the purpose of having them serve as suppliers of uniquely safe exchange media. In the limit, so far as transferable deposits are concerned, it would mean having a single, TBTF commercial-qua-central bank instead of today’s mix of TBTF and not-TBTF commercial banks. If that sounds like an improvement to you, you’re not thinking hard enough.

If they’re to avoid excessive risk taking, on the other hand, central banks can only manage to pay competitive returns on their ecash in one of two ways. They can operate so much more efficiently than commercial banks that they are able to more than compensate for their lower-yielding assets, or they can take advantage of their monopoly rents to subsidize their ecash business. The first possibility is far-fetched. The second isn’t. But as it amounts to a form of predatory pricing, the effect of which would be to drive central banks’ more efficient private rivals out of business, permitting it would be entirely contrary to the public’s welfare.[2]

If neither the safety nor the return on central-bank supplied ecash is likely to convince droves of bank depositors to switch to it, central banks might still encourage them by making their ecash easier to transact with than private substitutes. But this, too, is a tall order. Central banks have no experience in retail payments or in otherwise dealing with the general public: even the paper currency they produce is supplied to bankers only, who see to its retail distribution. Central bankers would therefore have to build their retail experience and facilities, whether online or brick-and-mortar, from scratch. In the meantime, they’d be competing head-on with commercial banks and other firms long and aggressively engaged in the business. Here again, the prospects for success seem dim, unless central banks resort to cross-subsidies to fund product-quality improvements, thereby gaining market share at taxpayers’ expense.

A Conflict of Interest

In suggesting that central bankers will find it difficult to out-compete commercial bankers unless they cross-subsidize their retail products, I am of course assuming that commercial banks and other private payment service providers will themselves remain as capable as ever of making their own products attractive to the public, by offering relatively attractive returns or otherwise.

Regulations can, however, severely limit the attractiveness of private monies, thereby making potential central-bank supplied ecash appear relatively more attractive. Examples of such regulations include high reserve requirements, other bank portfolio requirements, and usury laws. By making such regulations onerous enough, regulators could slant the digital-money playing field in central banks’ favor, thereby overcoming central bankers’ inherent disadvantages to usher in Sandbu’s ideal of a world in which central bank ecash is king.

But far from making Sandbu’s proposed reform appear more promising, the possibility in question supplies another reason for viewing it as a very bad idea. That’s because central bankers are among the regulators of private digital money suppliers. For that reason, allowing them to compete with such suppliers creates a conflict of interest, posing the risk that central banks’ regulatory actions will be influenced by their desire to preserve or enhance their share of the market for digital money.

Hello, Central Bank E‑Cash; Goodbye Payments Innovation

Finally, Sanbu, like many other boosters of central-bank ecash, blithely overlooks the chilling effect his proposed reform could have on future payments innovations. That we have private sector innovators to thank for the very existence of electronic money, starting with Western Union’s first telegraphic wire transfer in 1871, is (or ought to be) well known. We have them to thank as well for just about every other payments innovation, from checking accounts and lines of credit to ATMs, debit cards, PayPal, and cryptocurrency. For that matter, paper money itself appears to have been a private innovation, in China first of all, and much later in Europe, where London’s goldsmiths were issuing “running cash” notes more than a decade before the Bank of Sweden and Bank of England entered the market, which they later took over with the help of legislation that forced other banks to quit the business. How many of these private-market innovations would have happened had the innovators known that they were competing head-on with central bankers who might replicate their innovations whilst resorting to cross-subsidy financed predatory pricing to beat them at their own game?

There’s more than a little irony in proposals like Sandbu’s that would reward private sector payments innovators for their successful payments innovations by allowing central banks to employ those very innovations to assume a monopoly of retail payments. But irony is the least of it: the plan runs a very grave risk of putting the kibosh to future, desirable payments innovations. After all, once their monopolies of ecash are established, and assuming that they can resort to cross-subsidies to keep them, central banks will be under no competitive pressure to innovate. So while the prospect of their monopolizing retail payments today, using today’s leading-edge digital payments technology, may not seem all that unappealing, the prospect that they might go on employing roughly the same technology a century from now is considerably less so. Yet the possibility can’t be lightly set aside.

Paradoxically, appointing more innovation-inclined central bankers won’t necessarily help. Innovation is risky; indeed, it’s so risky that innovations fail more often than they succeed. When that happens in the private sector, the costs are born by the owners of the innovating firms. But when it happens in government (or quasi-government) agencies, taxpayers end up footing the bill.

All this is of course mere theory. But if you need empirical evidence, consider the U.S. Postal Service’s attempts at innovation, including its attempts to pioneer e‑mail. Perhaps central banks will somehow avoid the challenges that ultimately scuttled the USPS’s efforts. But I wouldn’t bet money on it.

__________________________

[1] Sandbu has since been joined at the FT by Martin Wolf, who first endorses Switzerland’s Vollgeld Initiative, and then suggests that allowing “every citizen to hold an account directly at the central banks” would work just as well. The Economist also endorsed the plan recently, prompting this rejoinder by Scott Sumner.

[2] The interest rate paid by the Fed on bank reserves has itself typically exceeded corresponding rates on short-term Treasury securities, thanks to its holdings of higher-yielding long-term securities, and hence to its having taken on considerable duration risk. Otherwise the Fed would presumably have had to subsidize those interest payments using seigniorage revenue from its currency monopoly.

[Cross-posted from Alt‑M.org]

In April 2018 Congressional Budget Office issued a report on the state of the federal budget in light of the “Tax Cuts and Jobs Act” enacted in December 2017, the “Bipartisan Budget Act” enacted in February 2018, and the “Consolidated Appropriations Act” enacted in March. The CBO estimates that, compared to the path it had previously estimated for the coming decade, the first of these will reduce federal tax revenues by 4% of GDP, while the latter two will add 2% of GDP to federal spending. In addition, debt service outlays are now expected to grow by 1.6% of GDP (rising to 3.0% from the current 1.4%). The cover of the CBO report featured the following chart with the stark caption: “Over the next decade, the gap between outlays and revenues is projected to be persistently large.”

The last recession officially ended in 2009. A persistently large deficit is remarkable when the country is so many years into a cyclical expansion, when tax revenues as a percent of GDP are ordinarily above their secular average. A traditionally prudent fiscal policy welcomes a falling ratio of debt to GDP during boom years to offset the debt growth of recession years, so that the debt remains stable over the business cycle as a whole, or even declines secularly between unusual spending spikes like wars. Today, however, the US federal government’s deficit is large and growing (and thus the US federal government’s debt is growing with unusual speed) despite an economy in its ninth year of expansion.

Economist John Cochrane on his blog recently reproduced a chart from a Deutsche Bank report (the report itself is behind a paywall), plotting the US unemployment rate (a rough proxy for the cyclical state of the economy and thereby federal tax revenue) against the US federal budget deficit. Usually the deficit and unemployment rate move together, because high unemployment (as in 2010) is symptomatic of a recession in real output, which also lowers federal tax revenues relative to federal spending. Well into an expansion (as in 2018) the reverse is true: unemployment is low in a recovered or booming economy, and deficits are low (or even surpluses arise) because tax revenues are cyclically high relative to expenditures. “Until now,” Cochrane comments.

The last two years have seen a widening of the deficit as a percent of GDP. As the chart above shows, the CBO expects this trend to continue. Today we have a growing economy with a low unemployment rate yet a growing federal budget deficit.

The CBO’s estimated “output gap” is a better proxy than the unemployment rate for the cyclical movement of income over the business cycle. (Here I put aside doubts about whether potential output is estimated in the best way, on the grounds that any shortcoming in the construction of the series is at least consistent over time.) Using the FRED database, I plot the federal budget deficit as a percent of GDP together with the output gap, measured as the percentage difference between the CBO-estimated “potential GDP” of the US economy and actual GDP. The overall pattern is that deficits follow the cyclical state of the economy.

There are two striking divergences from the pattern, however: (1) since the start of 2015, the deficit has been rising despite a falling output gap, and (2) between 1965 and 1968 under the Johnson administration the deficit grew despite a large negative output gap, meaning that output exceeded its estimated sustainable potential, which was made possible by an unemployment rate below its normal or natural rate.

As an aside, the chart also suggests that the budget surpluses of 1997–2000 were driven by a negative output gap (output unsustainably high). The explanation does not lie with any extraordinary efforts on the part of Congress or the White House beyond avoiding a repeat of the exceptional 1965–68 fiscal policies.

The Deutsche Bank report puts the 2018 US budget deficit at nearly 4.5% of GDP. This level is well above the level in any other major economy (see “Figure 6” in Cochrane’s blog entry). The Euro area as a whole has an average budget deficit of only about 1% of GDP. Even Italy is below 1.5%, while Germany runs a surplus. This is further evidence that the current US budget deficit path is abnormally – and avoidably — high.

A high deficit means a rapidly growing federal debt. Since 2007 Q2 the federal debt held by the public has more than doubled, to 75 percent from 34 percent of GDP. The stock of debt measured in dollars has tripled during this period, from just below $5 trillion to above $15 trillion. Thus far the global market has absorbed this outpouring of US Treasury bonds without requiring more than historically low yields on the bonds. But how much larger can the federal debt grow before it causes obvious trouble? The CBO projects that the debt held by the public will rise to 100 percent of GDP by 2028. The CBO’s director Keith Hall has darkly remarked of his agency’s projections: “Such high and rising debt would have serious negative consequences for the budget and the nation. In particular, the likelihood of a fiscal crisis in the United States would increase.”

What would a “fiscal crisis” mean for the United States? First, consider a country that borrows in a currency that it cannot print, for example Greece (in euros) or Ecuador (in dollars). Such a country faces a fiscal or sovereign debt crisis when a mounting expectation of default (inability or unwillingness to repay all the bondholders in full and on time) drives up the default risk premium part of the real yield that bondholders require to buy its bonds. With real bond yields above the economy’s real growth rate, the debt compounds faster than GDP, boosting the ratio of debt to GDP, which raises the default risk premium still further, and so on. Without a rescue, this culminates in actual default (Ecuador 2008, Greece 2011).

The United States, of course, issues bonds in a currency that it can print in unlimited amounts, so there never need be a failure to pay the number of dollars owed, even if tax revenues are insufficient to cover all obligations. Mounting debt therefore raises not the risk of default risk in the conventional sense but the risk of repayment in increasingly cheapened dollars. To avoid Treasury default the Fed can pursue a policy of rapid money-printing that raises the inflation trend. Potential buyers of unindexed bonds, who can be expected to become concerned about this possibility even before inflation rises, will add an inflation-risk premium rather than a default-risk premium to real bond yields. As Cochrane comments, “Borrowing in your own currency only means that our government can substitute inflation and devaluation for explicit default, if it refuses to fix its finances.”

The concept of the Fed needing to “print money” (expand the monetary base) to allow the Treasury to service its bond debt was discussed decades ago in connection with what Sargent and Wallace called “unpleasant monetarist arithmetic.” The key concept they introduced was that simply rolling over a country’s debt, once the Treasury bond yield exceeds the economy’s growth rate, implies an endlessly growing ratio of Treasury debt to GDP. At some high debt-to-GDP ratio the market becomes saturated with Treasury bonds, and bond revenue reaches a maximum. This means that further borrowing raises the yield on the entire debt (as it rolls over) so much that the new borrowing provides no net revenue. There is a peak to the bond-revenue “Laffer Curve.” If spending and tax commitments mean that the deficit cannot be reduced (“austerity” is politically impossible, including sizable government asset sales), and no stream of external gifts can be obtained, the last resort for avoiding default is to increase the revenue from printing money. Growth in the monetary base must rise to help finance the mounting debt, given the other policy constraints.

Before the Fed began in 2008 to pay interest on part of bank reserves (namely banks’ deposits held at the Fed), there was a clear distinction between the base and the interest-bearing debt of the federal government. With banks not incentivized to hold excess reserves, higher base growth translated into higher growth in the broader money stock in circulation, say M2. Faster money growth in turn meant more units of money chasing each unit of goods and services, causing faster price level growth. Domestic money-holders bore the burden of a higher inflation tax.

Since gaining authority to pay interest on reserves in 2008, by contrast, the Fed has raised the interest rate it pays on excess reserves as necessary to persuade commercial banks to hold them rather than to create additional deposits by “lending out” the excess reserves, and thus has kept M2 growth from rising with higher base money growth. Thus the dramatic base growth since 2008 has not yet imposed an inflation tax. Instead, insofar as the Fed has acquired Treasury rather than private securities, and has paid banks the full Treasury yield rate on their depository reserves, it has been meant merely a fiscally pointless swap of one interest-bearing government liability in the hand of the non-Fed public (commercial bank deposits on the Fed’s books) for another (Treasuries held outside the Fed). But this means that Fed purchases of Treasuries no longer in effect retire them. That is, Fed acquisition of Treasuries no longer reduces the consolidated Fed-Treasury burden of debt service to the public The Treasuries held by the Fed are now properly counted (as they have been all along) as federal debt held by the public. Rising Fed holdings of Treasury debt, financed by interest-bearing reserves, are part of the rising ratio of debt to GDP.

The fiscal status quo ante would be effectively restored if Congress would take up George Selgin’s proposal that the Fed pay interest only on required reserves, not on excess reserves. Holding the interest rate on reserves constant (at zero or even above), faster base growth would again mean higher inflation. But isn’t the Fed committed to not exceeding a 2 percent rate of inflation? Only rather weakly. It is a commitment that the Fed adopted at its own initiative, and can repudiate at its own initiative. It is not embodied in any binding legislated mandate. Ask yourself: if it came down to a choice between the Treasury defaulting on bonds because permanently faster money-printing would permanently raise inflation above 2 percent, and raising the Fed’s current inflation-rate target, which option do you think would prevail? (I would expect higher inflation, though Jeff Hummel has made a coherent case for expecting default.)

It is to avoid this kind of inflationary “solution” to mounting US federal debt that we dollar-holders must be concerned to do something about the current trajectory of high US budget deficits and a persistently rising ratio of debt to GDP.

As all dedicated Alt-M readers know, I am a big fan of the Canadian banking and monetary system that flourished between Canada's Confederation in 1867 and the outbreak of the First World War.  Besides thinking it was a darn good system, I also regard it as the best example, together with Scottish banking during the first half of the 19th century, of a "free" (that is, largely unregulated) banking system.

But was the pre-WWI Canadian banking system really a "free" banking system? Although Canada's commercial banks were free from many sorts of regulatory restrictions imposed on banks elsewhere, during its "free banking" era, all of Canada's banks were "chartered" banks, where, as Stephen Williamson observed in a recent twitter exchange, Parliament had to approve each charter.  "To me," Stephen wrote, "that seems like a severe entry barrier."

And so it may seem to most people. Yet a closer look at the facts suggests a different and changing picture. Despite banks' need to secure Parliamentary charters, entry into Canadian banking was relatively free until at least 1890, when stiffer bank capital requirements made it harder for prospective new entrants to qualify for a charter. Canadian bank mergers were in turn made easier by the Bank Act of 1900. As a result of these two changes the number of Canadian banks fell dramatically over the course of the first quarter of the 20th century, making it increasingly less appropriate to treat Canada's banking system as an example of "free" (in the sense of "laissez-faire") banking.

Charters before and After Confederation

Although prior to Confederation the chartering of any sort of corporation in Canada was technically a Royal Prerogative, British authorities were generally inclined to ratify charters granted to banking companies by Canadian authorities, provided those charters conformed to certain regulatory provisions the British authorities favored (see Shearer 2005). But it wasn't until the 1840 Act of Union, uniting the colonies of Upper and Lower Canada into a single Province of Canada, that uniform regulatory provisions were first settled upon, which were to be imposed both on newly-formed banks and on established  ones as conditions for having their charters renewed.

It was while these terms were in effect, in 1854, that the Canadian-American Reciprocity Treaty was signed, giving a big boost to Canadian trade, and generating a corresponding increase in applications for bank charters. According to Byron Walker, General Manager of the Canadian Bank of Commerce and former president of the Canadian Bankers Association, between 1855 and 1866, inclusive, "more than twenty-five charters were granted," with fifteen new banks actually coming into business.

Confederation in 1867 brought with it the necessity of framing a general bank act for the newly established Dominion of Canada, a version of which was adopted in 1870, and then revised in 1871. The revision made the Canadian Parliament responsible for both the granting and decennial renewal of bank charters.

The necessity of a charter certainly distinguished the Canadian system from its Scottish counterpart, for although three chartered banks took part in the latter system, it also included over 30 unchartered banks, most of which issued notes, and many of which were joint-stock companies. However, as Walker explains in his 1896 History of Banking in Canada,  the distinction wasn't as important as one might think:

In Canada, we have had one section of the people who have been so enamored of freedom that they have desired to see banking as well as other privileges reduced to the mere necessity of applying for incorporation under a general act, together with subscription of the smallest amount of capital which seemed possible to propose. But, as a rule, people of British origin want merely all the liberty which is compatible with freedom from license. So that while, in the main, Parliament has clung to its prerogative of refusing a charter if it chose to do so, during 50 years at least, it would not have dared to exercise the power except in the event of a clearly fraudulent application for a charter. Nor would it dare, although it has the power, to give special privileges to any one bank.

To drive home the point, Walker compares the Canadian situation to that of the U.S., where the National Bank Acts allowed banks to be established anywhere without need for a special charter.

In the United States, a certain number of individuals having complied with certain requirements -- more numerous and complicated, by the way, than the Canadian requirements -- became thereby an incorporated bank, if we regard the consent of the Comptroller of the Currency as a matter of form. In Canada, when a certain number of individuals have complied with certain requirements, they are supposed to have applied for a charter, which Parliament theoretically might refuse, but which, as a matter of fact, would not be refused unless doubt existed as to the bona fide character of the proposed bank. Then, as in the United States, on complying with certain other requirements and obtaining consent of the Treasury Board (performing in this case the same function as the Comptroller of Currency in the United States), the bank is ready for business.

Other Perspectives

Readers may well doubt that we should take an established Canadian banker's word for it when he tells us that entry into banking in Canada was practically as open as it was in the U.S. But Byron Walker was hardly the only person who thought so. Here, for example, is the opinion of R.M. Breckenridge, a highly regarded historian of Canadian banking (he was commissioned to write the National Monetary Commission volume on The History of Banking in Canada) who, though a Canadian and a businessman, was not himself a banker. In Breckenridge's 1895 book on The Canadian Banking System, he says that

The disposal of bank charters, has never been marked by the fraud or partizanship which make the record of some of the American commonwealths so discreditable in this respect... . Yet charters have been easily obtained, too easily obtained. Since confederation forty-four charters have been granted, and only five proposed charters reported on adversely... . Twenty of the forty-four have been forfeited for non-user. ... Any new bank may now be chartered as soon as the projectors convince the disinterested committee of Ministers and heads of departments known as the Treasury Board, that their intentions are honest and that they have financial backing. A favorable report by the Treasury Board or the House Committee on Banking and Currency makes the bill a Government measure and ensures its passing. The Canadian banks have enjoyed no monopoly against the entrance of new competitors bona fide  into banking, nor have the shareholders profited from investments in stocks which others might not obtain.

Writing in 1910, when (as we shall see), the Canadian banking system had already begun a process of consolidation, American economist Joseph French Johnson (1910, p. 86) observed, in his own National Monetary Commission volume, that, although "Some critics insist that the difficulties in the way of establishing a new bank are so great that the existing banks practically possess a monopoly," the criticism was still "not altogether justifiable." "A dozen or more new banks have been chartered since 1890," Johnson observed, "and there is no evidence whatever that Parliament has refused a charter to any set of deserving incorporators."

The Beginnings of the End: Canadian Banking Consolidation

But while entry into the Canadian banking system may be said to have long been relatively free when Walker, Breckenridge, and Johnson wrote about it, a sea change was underway. The change was a result of two modifications to Canada's banking laws. The first was the implementation of stricter minimum capital requirements included in the the Bank Act of 1890. That provision required, not only that banks meet a $500,000 capital requirement, but that that capital be fully subscribed within a year of the granting of the charter, and that at least one-half of those subscriptions be paid in, before the Treasury Board would grant the bank a certificate authorizing it to commence business.

These provisions marked a substantial change from those of the previous Bank Acts. Although the 1871 Act had also required banks to have at least $500,000 in subscribed start-up capital, only $100,000 of that amount had to be paid up. Moreover, as John Turley-Ewert points out in his excellent dissertation on the history of the Canadian Bankers Association (p. 14, n 55), because government authorities never bothered to enforce the provisions, in practice capital requirements appear to have been "meagre rather than substantial," and especially so since the 1871 Act imposed no requirements at all on private bankers, who therefore "successfully competed with the chartered banks through the 1870s and 1880s." Besides raising the paid-in component of bank capital requirements, the 1890 Act also sought to give the law some real teeth.

Despite one notorious failure, involving the Ontario-based Farmers of Canada,[1] the new law appears to have largely achieved its intended purpose of limiting bank entry. Breckenridge himself understood the new capital requirement to be

a legal step in the direction of making the organization of new banks more difficult. No new bank has entered the field since 1885. ...Of the fourteen banks chartered in 1883 to 1893 inclusive, only five could comply with the requirements of the Bank Act and actually began business; three of the five have already been put in liquidation, two in 1887 and one in 1893. It may be expected that hereafter both people and Parliament will be disposed closely to scrutinize applications for new charters. The enthusiasm for new banks...has abated... . The tendency of the number of banks to remain stationary, or even to diminish, so pronounced in English and Scotch banking,[2] is thought a factor of considerable influence in the Canadian situation.

The second modification, included in the Bank Act of 1900, was in itself a step in the direction of more rather than less freedom in banking, consisting of the relaxation of what had been a very strict policy regarding Canadian bank mergers. Before 1900, it took an act of Parliament to authorize a merger, just as it took such an act to grant charters to new banks; Parliament was not, however, inclined to approve of mergers. The 1900 revision allowed any bank to acquire another bank's assets without Parliamentary action provided the merger was approved by the Governor in Council pursuant to a recommendation of the Treasury Board. Consequently mergers become much more common: whereas from Confederation until 1900 only seven bank mergers occurred, the period between 1900 and 1926 witnessed another 27.

Because of these changes, and as Breckenridge had anticipated, the number of Canadian banks fell rapidly. In his 1929 book, The Banking System of Canada, Benjamin Beckhart observed that "The tendency toward fewer and larger banks has been perhaps the outstanding characteristic of the banking structure of Canada within recent years." While the number of Canadian banks peaked at 41 in 1885-6, Canada still had 38 banks a decade later, when Walker and Breckenridge penned their respective histories. In 1910, when Johnson wrote, the number had fallen to 29; and he had no doubt that it would fall further:

The tendency in Canadian banking...is unmistakably toward combination. The chartered banks possess no monopoly now, but the situation is such that the large banks have a great advantage over the small ones and seem destined to get most of the new business that will be created in Canada in coming years. Their prestige assures their branches are welcome in every new community. No business is too large for their resources, and none is so small as to be despised. They are able to open more new branches than their small competitors, and can much better afford, if need be, to operate them for a time at a loss. It will not be surprising, therefore, if Canada has fewer banks ten years hence than now (Johnson 1910, pp. 134-35).

By 1914 there were, in fact, just 22 chartered banks; and by 1935, when the Bank of Canada was established and granted a monopoly of note issue, only 10 were left. Because the decline reflected both a lack of new entrants and mergers, it brought with it a greater concentration of assets in a handful of banks, with the four largest in 1929 holding 77% of the total (Beckhart p. 328).

In short, by the time the Bank of Canada took over, Canada's private currency system could no longer be regarded as an essentially "free" system, differing only in degree from the Scottish system in its pre-1845 heyday. For this reason Canada's era of (relatively) free banking is best regarded as lasting only until 1914 at the latest, when consolidation, though already underway, had as yet left the system's operating principles fundamentally unchanged.

Banking without Sin

What ultimately doomed free banking in Canada, even before the advent of the Bank of Canada,  was Canadian authorities' growing unwillingness to tolerate any risk of depositor or banknote-holder losses in connection with bank failures, even when those failures posed no macroeconomic threat. The government may first have hoped that stiffer capital requirements would rule-out losses stemming from bank failures. But when three banks failed between 1890 and 1900, and two of those failures -- those of the Banque Du Peuple in 1895 and of the Banque Ville Marie, in 1899 -- resulted in substantial depositor losses, its hopes were disappointed. The government's subsequent decision to encourage bank mergers was supposed to serve as an alternative and more palatable means for winding-up failing banks. Until the advent of deposit insurance many decades later the combined effect of these steps was to all but guarantee a gradual move towards banking-industry oligopoly.

The lesson here seems obvious enough. In banking as in other enterprises, genuine competition is a matter of creative destruction, where the "creative" part calls for relative ease of entry,  while "destruction" means that badly managed banks are allowed to fail, even if that means that creditors sometimes bear losses. The occasional liquidation of individual banking firms, and the losses suffered by those firms' shareholders and creditors, are part of the cost of having a robust and stable yet nonetheless dynamic banking system. If they go too far in their attempts to avoid such costs, regulatory authorities could end up undermining, rather than bolstering, the banking system's long-term integrity.

As the late Allan Meltzer remarked, “Capitalism without failure is like religion without sin.” For a quarter century or so, Canada's banking authorities left Canada's banks and bank customers to the mercies of the old-time capitalist religion. By so doing they allowed one of the world's most brilliant banking systems ever to flourish, albeit at the cost of exposing Canada's bank patrons, not only to bank failures, but to occasional banking swindles. But Canadian authorities eventually found the trade-off impossible to tolerate, preferring to stamp out sin altogether, even if that meant transforming Canada's competitive though imperfect banking system into an immaculate banking cartel.

__________________
[1]Established for the express purpose of advancing credit to farmers on more favorable terms than those offered by existing chartered banks, the Farmers Bank struggled to raise the necessary capital. Wishing to see the bank succeed, William Fielding, Canada's Liberal Finance Minister, first granted its promoters several extensions, and then, in 1906, instructed the Treasury Board to issue the bank a certificate to begin business solely on the basis of a sworn oath to the effect that the necessary capital had been raised, by the bank's general manager, W. R. Travers, despite having been warned  by persons in a position to know that Travers' claim was false. The Farmers Bank suspended payments on December 19, 1910, its failure having been due, according to a royal commission, to its management's "gross extravagance, recklessness, incompetency, dishonesty, and fraud." Although that same commission absolved Fielding of negligence, Travers was eventually found guilty of having violated the Bank Act, for which he was sentenced to six years in prison.

[2]The consolidation referred to here was a consequence of the passage of Peel's Bank Charter Acts of 1844 and 1845, which effectively ended new entry into the banknote issuing business in the United Kingdom.

__________________

More from Alt-M on Canadian free banking:

• Wrong Lessons From Canada's Private Currency, Part 1
• Wrong Lessons From Canada's Private Currency, Part 2
• Wrong Lessons From Canada's Private Currency, Part 3
• Did Free Banking Stabilize Canadian NGDP?
• There Was No Place Like Canada
• That's Right: "Famously Sound and Famously Stable"