From Ricardo to Sraffa: A Quest for a Modern Classical Standpoint on Money

Abstract

The topic of the chapter is the possibility of a consistency between Sraffa’s price system and Ricardo’s theory of money. Some directions in which such consistency has been looked for are discussed, focusing on a specific one based on my understanding of Ricardo’s theory of money (in Deleplace, Ricardo on Money, A Reappraisal, Abingdon: Routledge, 2017) and on some writings by Sraffa prior to Production of Commodities. As such, the chapter aims at contributing to a quest for a modern Classical standpoint on money, if by “modern Classical” one means faithful to Ricardo—the greatest Classical author on money—and consistent with a Sraffa system of prices.

Appendices

Comment

1.1 A Comment on ‘A Quest for a Modern Classical Standpoint on Money’

• Carlo Benetti &
• Jean Cartelier 

The subtitle of GD’s chapter is explicit: “A quest for a modern Classical standpoint on money”; while wishing to be “faithful to Ricardo”, GD’s aim is to be “compatible with Sraffa’s price system ”. Benetti and Cartelier (2020) chose to favour the second objective at the cost of the first one. The coherence of Sraffa’s price system with Ricardo’s monetary theory, as interpreted by GD, raises difficult problems.

GD’s central statement is that “the adoption of Sraffa’s prices by a monetary economy, in accordance with Ricardo’s theory of the currency, requires a specific working of the market for the commodity selected as standard of money .” The recent, excellent book published by GD Ricardo on Money reinforces this thesis. However, two points need to be raised: the specificity of the gold market and its relation to Sraffa’s natural prices .

It is true that, according to Ricardo, the market for the standard of money works in a specific way. But the real distinctive feature of the gold standard is its legal price in terms of £, the consequence of which is that, in a closed economy, the market price of gold cannot diverge from the legal price. On the other hand, in GD’s construction of a “modern Classical standpoint on money” gold is assumed to be produced outside the economy. As a result, it becomes problematic to link gold with the domestic production system determining natural prices and the rate of profit . The project of being “consistent with a Sraffa system of prices” is seriously impacted.

1.2 The Specificity of the Gold Market

GD states that, in the Classical monetary theory (i.e. inspired by Ricardo), the suitable amount of convertible notes issued by the Bank is regulated by the arbitrage between the market price and the legal price of gold . This proposition will be discussed after a brief summary of GD’s analysis.

1.2.1 Smith-Thornton’s Arbitrage

In Benetti and Cartelier (2020, § 2) we showed that in an (abstract) economy with convertible banknotes , the price of gold in the domestic market cannot differ from the legal price and any possible gap between these prices quickly vanishes.¹³ GD’s main critique consists in an argument taken by Ricardo from Thornton who, in turn, was inspired by Smith (1776 ) . The reference text is found in High Price of Bullion, p. 58–9 where Ricardo presents Thornton’s analysis as a possible alternative to the regulation of the quantity of money by means of the exportation of gold which he has just analysed in detail (1810–1811, pp. 54–8, see also our paper, pp. 252–3). According to Thornton, in order to increase an insufficient stock of gold, the Bank of issue¹⁴ purchases gold bullion in the domestic market, although the market price is higher than the legal price at which individuals obtain gold against their notes. Individual gains are the Bank’s losses. The whole argument is taken literally by Ricardo as showed by Thornton’s long quotation (1810–1811, pp. 58–9), and endorsed by GD: “The bank thus provides the missing side of the market for bullion, in the absence of any demand by all other agents who may obtain gold from the bank at the legal price” (p. 504).

In the gold market , as Thornton notices, “The one party [the individuals] will be melting and selling while the other [the Bank] is buying and coining”. The objects traded on this market are also heterogeneous. Individuals sell a commodity against £. The Bank does not buy a luxury good. It buys gold as the material of coins and the basis of issuing banknotes .

The Bank inevitably suffers losses. This is why Thornton speaks about a “very unequal war” so that the Bank “will be likely to be tired sooner than its adversaries”. It is Ricardo who sets out the true solution (mentioned by GD p. 506): the Bank “would be obliged ultimately […to…] withdraw part of their notes from circulation, till they should have increased the value of the remainder to that of gold bullion ”. As in the case of the exportation of gold, it is the only sensible measure to reach the desired value of the ratio between the quantity of banknotes and the Bank’s stock of gold (we dubbed μ).

1.2.2 Comments

It should be emphasized that our only aim is to clarify the notion of arbitrage between two prices when one of them is fixed by law.¹⁵ The main theoretical weakness of Smith-Thornton’s argument is that the losses of the Bank result from its odd monetary policy . For the same amount of £ received by the Bank against gold and spent to redeem gold in the market, the quantity of gold in the Bank’s vault decreases; for the same quantity of gold in the vault of the Bank, the amount of circulating banknotes increases. In both cases, multiplier μ increases. The “Penelope effect ” as called by GD,¹⁶ implies that Penelope undid in the night more than what she did in the day. It is easy to guess the end of such a tale.

The difficulty does not lie in the Bank’s losses as such. Depending on circumstances they could be justified. The problem is that these losses are the direct consequence of the Bank monetary policy which is diametrically opposed to the very aim of the Bank i.e. to decrease the ratio μ. Why does the Bank provide “the missing side of the market for bullion” (p. 504) whereas this action worsens the problem that motivated its decision to act? GD states that “when the metallic reserve of the bank of issue falls below a certain limit, the bank is forced (ital. added) to purchase bullion in the market, although by doing so it has to pay a higher price than the legal price at which it gives gold against its notes”. Such an assertion is not acceptable: the insufficient gold reserve cannot “force” the Bank to apply a monetary policy that reduces it further. The constraint on the Bank is quite different. As stated by Ricardo, the Bank ends up being “obliged” (1810–1811, p. 59) to reduce the amount of circulating notes, which is the only way to bring μ down.

Finally, why does not the Bank adopt this measure from the beginning? We do not see how to justify the idea that the initial reserve of gold is both too low—so that the Bank is motivated to buy gold at a price higher than the legal price—but not low enough to force the Bank to adopt the only rational action that is the withdrawal of “part of their notes from circulation” as stated by Ricardo. Why must the initial stock of gold be reduced further?

In our paper we argued that there is no arbitrage between two prices when one of them is legally fixed. We now complete our argument: the supply of gold in the domestic market is positive only if the market price is higher than the legal price; but in such a case the demand for gold is also positive only if there is at least one agent whose economic behaviour is incoherent. As rightly stated by GD, our position “contradicts” the “Penelope effect ” he qualifies as “an important argument.” We have just showed that such a contradiction is in favour of our position and not against it as GD asserts.¹⁷

Besides, it seems to us that Ricardo clearly favours the regulation of the quantity of money through the export of gold by individuals. Gold is bought by foreigners and the role of the Bank would be to regulate the issuance of money and not to act in the market instead of individuals, as stated by GD. A deep doctrinal reason is suggested by Ricardo (1810–1811, pp. 55–6): “The exportation of the specie may at all times be safely left to the discretion of individuals; it will not be exported more than any other commodity, unless its exportation should be advantageous to the country. If it be advantageous to export it, no laws can effectually prevent its exportation. Happily in this case, as well as in most others in commerce where there is free competition, the interests of the individual and that of the community are never at variance.”

1.3 What About a Relation with Sraffa’s Natural Prices?

Two points will be made thereafter. GD’s requirement that gold be produced outside the economy (i) is not well-founded since there is in fact no contradiction between the legal price of gold (in £) and the natural price of gold (in a single or composite commodity ) when gold is domestically produced; and (ii) is not appropriate for an integration, intended by GD, of Sraffa’s natural prices in Ricardo’s monetary theory.

1.3.1 No Contradiction to be Found When Gold is Domestically Produced

In Benetti and Cartelier (2020), we also attempt at exploring the spiny question of a possible link between Sraffa’s system of prices , on the one hand, and the Classical thought about money and market prices , on the other. What follows is tentative and is put forth for the sake of discussion. GD maintains that compatibility between Sraffa’s theory of natural prices and Classical theory of money requires that gold (the standard of money ) must be produced outside the domestic system of production. GD’s assertion is surprising. Not as a general principle since anthropologists have shown that in many societies money effectively comes from abroad, but as a specific feature of Ricardo’s views. A straight consequence of that requirement is that the natural price of gold is no longer determined by the conditions of production of the economy under consideration.

Why does the integration of money in Sraffa’s system of natural prices require the expulsion of gold production away from the economy? This is in fact related to a wall separating money and prices. According to GD, the market price of gold is subject to “two apparently contradictory adjustment processes”, gravitation around Sraffa’s natural price , on the one hand, regulation through the legal price, on the other. “A conclusion emerges: this double regulation of the market price of gold bullion may only be non-contradictory if one assumes that gold bullion is produced outside of the country in which it is used as the standard of money” (Deleplace 2017, pp. 272 and 273; Deleplace’s emphasis).

In Benetti and Cartelier (2020) the same problem occurs but we have not found any contradiction. Let us start from the following system of natural prices determination à la Sraffa (system (2*) in Benetti and Cartelier 2020):

$$ {\displaystyle \begin{array}{l}\left(1+r\right){A}^{\prime }p=p\\ {}\left(1+r\right)\lambda \sum \limits_{i\ne g}{a}_{ig}{p}_i=\lambda {p}_g\\ {}{p}_g=1\end{array}} $$

(2*)

where r is the rate of profit , A ^′ the transpose of A the square matrix of the a _ij’s, a _ij the quantity of commodity i used for the production of commodity j, and p the vector of natural prices .

The second line of (2*), which concerns gold (subscript g), is formally identical to the others except for parameter λ which applies to the natural quantity of gold (equal to one as for any other commodity). Gold is supposed to be a non-basic commodity in accordance with GD’s position, which means that its conditions of production do not determine the rate of profit and the natural prices of basic commodities. Once the system of production of basic commodities has been solved (first line of (2*)) and when market prices have converged toward natural prices , as GD seems to take for granted, it is easy to get money (natural and market ) prices by using the legal definition of the £, that is:

$$ {p}^{\hbox{\pounds}}={p}^g\frac{1}{\beta}\hbox{\pounds} $$

(3*)

where β is the quantity of gold which defines the £ and p ^g is the vector of natural prices —solution of (2*)—expressed in gold.

If “a mine of gold were discovered” (Ricardo 1810–1811, p. 54) and if the technique did not change, parameter λ would increase but not the “wants of commerce ” (see Benetti and Cartelier 2020). Therefore no variation of the quantity of means of payment would take place. The additional quantity of gold is used as a commodity by contrast with what happens according to Ricardo who considers that the total quantity of gold is coined.

If the relative value of gold decreased as an effect of a better production technique, vector p ^g would be multiplied by a scalar greater than 1. The same scalar would also apply to the prices expressed in £, which are nothing but natural prices in gold multiplied by constant $\frac{1}{\beta }.$ The “wants of commerce ” increase and are met by the issuance of additional means of payment as a counterpart of additional loans (see Benetti and Cartelier 2020 , p. 247)

1.3.2 Some Puzzling Consequences of Gold Being Produced Abroad

Requirement that gold has to be produced abroad does not seem to affect GD’s reasoning. This is the consequence of his neglect of the general problem of gravitation of market prices around natural prices . His focus is on the specific question of the gold market only. However, elaborating a coherent theory embracing both Sraffa’s natural price system and Classical theory of money cannot avoid considering market prices in general. This is what GD implicitly comply with in his book where market prices are present under the disguise of the so-called “wants of commerce ”. It is not the proper place for developing this difficult point. We shall content ourselves with insisting on the necessity to inquire into the relation between natural and market price in the specific case of gold (the market price of gold being its legal price).

Two formal expressions of gold being produced abroad may be considered.

In the first one, the second line of system (2*) is cancelled which amounts to eliminating the natural price of gold . However, the natural price of gold (produced abroad) enters his equation (5.8) (in Deleplace 2017, p. 187), which gives the condition for importing gold . Gold imports are subject to a comparison between the cost of buying gold abroad (including the domestic natural rate of profit ) and its legal price. He determines the cost of importing gold as its natural price abroad (assumed to be equal to its market price ) corrected by the rate of exchange times (1 + r) given by (2*). That importing gold from abroad makes sense only if the cost of it is inferior or equal to its legal price is pretty sure. Pretty sure also is the fact that the natural price of gold does not play any role here the domestic natural price of gold has disappeared with the cancellation of the second line of (2*) and the natural price abroad is replaced by its market price as a consequence of gravitation . Note that GD also requires that gold is not the standard of the monetary system abroad. This last consideration shows that GD cannot get rid of the embarrassing question of gravitation , at least for gold produced abroad, gold not being the standard there. Why should the market price of gold be equal to its natural price if no gravitation mechanism is considered?

Notwithstanding a further point. GD’s equation (5.8) has a clear general signification but we have to wonder who is concerned by it. Does it mean that a special branch or industry runs gold external trade? If not so, why invoke the domestic natural rate of profit ? Any industry (corn, iron, etc.) could get an extra profit (in addition to its natural one) by importing gold if condition alluded above (equation 5.8) is met even if this extra profit is small.

A second interpretation has to be explored in order to escape from the critique above. Let us consider the complete system (2*) with its second line. Instead of production stricto sensu, that second line would show the gold external trade as a special branch. Such a view seems perfectly compatible with GD’s requirement that gold be a non-basic commodity produced abroad. Inputs a _ig would be interpreted as the elements of the vector of commodities bartered against gold, imported gold being viewed as the specific output of that branch subject to the rule of the uniform rate of profit . Then, it could be possible to get the (domestic) natural price of gold given by (2*). It would be the result of inserting in the second line of (2*) a _ig’s natural prices and the rate of profit which are solutions of the first line of (2*). The last line of (2*) allows getting natural prices expressed in gold.

Now we face two natural prices (domestic and foreign) of gold. But they do not play any role in GD’s equation (5.8) since only the foreign market price of gold and the domestic legal price of gold are relevant (with the rate of exchange).

It is not possible to go further unless we would be obliged to embark upon a theory of international natural prices which is beyond the scope of these limited comments. It is enough to be aware of the numerous and difficult questions that any Classical theoretician of money has to overcome in order to make that theory consistent with a Sraffa system of prices.

Another problem comes from the market price of gold . In a closed economy, the market price of gold is necessarily equal to its legal price (Benetti and Cartelier 2020). What can be said about it when gold is imported?

For Classical economists, since Smith’s Wealth of Nations chapter vii, “Cantillon’s rule” determines market prices in general. That rule applies here in a specific way. Let us imagine that gold external trade branch, as a special industry, borrows a given amount of $ from the Bank abroad. Thanks to these $, it buys gold abroad at market price . Assume that quantity of imported gold, when brought to the Bank of England , allows that industry getting a sufficient amount of £ and allows paying back the foreign Bank after conversion in $ (we abstain from any consideration about exchange rate for the sake of simplicity). The foreign market price of gold is given, according to Cantillon’s rule, by the ratio of the total of means of payments ($ and £ made homogenous thanks to the rate of exchange) divided by the quantity of gold brought to market. That quantity and the total amount of means of payment are not known. Even if this does not affect the condition alluded by GD (equation 5.8), it shows how long is the way toward an integration of a Classical theory of money —not separable from market prices determination—in Sraffa’s natural prices theory .

Reply

1.1 A Reply to Benetti and Cartelier

• Ghislain Deleplace 

1.2 The Compatibility between Sraffa’s System of Prices and Ricardo’s Theory of Money

The common issue under discussion in my chapter, in Benetti and Cartelier’s comments on it, and in the present reply, is the compatibility between Sraffa’s system of prices and Ricardo’s theory of money. Consequently what are left aside pertain to Sraffa’s own monetary views before Production of Commodities by Means of Commodities—tackled in Sect. 3 of the chapter—and to Ricardo’s monetary views not directly linked to his theory of prices—such as the detail of his stipulations about the best monetary system. This discussion, however, cannot be freed from two other references by the same participants, which have a distinct object in their own right. One is my book Ricardo on Money. A Reappraisal: although it claims that it may contribute to the integration of money in the modern (Sraffa’s) theory of Classical prices (Deleplace 2017, pp. 384–6), it pursues another goal, which is to provide a coherent reconstruction of Ricardo’s theory of money. The second reference is a recent paper by Benetti and Cartelier entitled “From Ricardo to Sraffa: Gold as Monetary Standard in a Classical Theory of Money”: the authors explicitly warn that “to establish ‘what Ricardo has really said’ is not our topic. We have preferred to inquire into what could or should be a Classical theory of money and prices sixty years after Sraffa’s Production of Commodities by Means of Commodities publication.” (Benetti and Cartelier 2020, p. 241). One difficulty of the discussion is that some of the arguments exchanged cannot be fully understood without referring to these two other publications.

An interesting aspect of these two references is that they both offer a more or less detailed formalization. This should help locating precisely on which points there is a disagreement between studies which have in common to praise Sraffa’s contribution to the Classical theory of prices and to reject the usual understanding of Ricardo’s monetary theory in terms of the Quantity Theory of Money . The model of Ricardo’s theory of money developed in Deleplace (2017) leads to three main results:¹⁸ (a) The value of money varies positively with the natural price of the standard and negatively with its market price (what I call “the Money-Standard Equation”); (b) The standard (e.g. gold) is necessarily exterior to the system of production that determines the natural prices of the commodities in the economy where the standard-based currency circulates; and (c) When the quantity of money is in excess of what is required by “the wants of commerce ”, it is endogenously adjusted through a contraction of the note issue that is forced by the unfortunate consequences for the issuing bank of the rise in the market price of the standard (what I call “the Penelope effect”).

In their comments, Benetti and Cartelier target result (b): “GD’s requirement that gold be produced outside the economy (i) is not well-founded since there is in fact no contradiction between gold legal price (in £) and gold natural price (in a single or composite commodity ) when gold is domestically produced; and (ii) is not appropriate for an integration, intended by GD, of Sraffa’s natural prices in Ricardo’s monetary theory” (p. 519). They also target result (c): “As rightly stated by GD, our position ‘contradicts’ the ‘Penelope effect’ he qualifies as ‘an important argument’” (p. 518). Benetti and Cartelier’s assumption that under convertibility the market price of gold cannot diverge from the legal price also amounts to rejecting result (a) since the value of money does not vary with a market price of the standard that they assume to be fixed. This assumption of a fixed market price of gold is central in the alternative Classical model of a standard-based money in a Sraffian system of prices which they propose in Benetti and Cartelier (2020).

In the same order as they appear in Benetti and Cartelier’s comments, I will discuss below their critiques of my results (c) and (b), before emphasising that their fixity assumption is inappropriate in Ricardo’s theory of money, hence in any attempt at making it compatible with Sraffa’s system of prices. But before that, a caveat may help the reader. In accordance with their commitment in their 2020 paper to leave aside “what Ricardo has really said”, Benetti and Cartelier do not suggest that my reconstruction of Ricardo’s theory of money might be unfaithful. Consequently, the critiques they raise at me should be considered as applying to Ricardo himself. Besides, my object has been to reconstruct Ricardo’s theory of money in a way that makes it: (1) consistent with the whole of Ricardo’s writings on money; (2) internally coherent; and (3) compatible with Sraffa’s system of prices. It has not been to contend that this reconstruction is the only (and even the best) way to analyse a monetary economy—an assertion which would be beyond the point.

1.3 The Divergence of the Market Price of the Standard from its Legal Price

Two questions may be asked about this divergence : can it exist, and if yes, what happens to it? I have shown elsewhere (see Deleplace 2020) that, even if they may disagree on the answer to the second question, the (minority) unorthodox scholars on Ricardo’s theory of money all answer to the first question in the positive and consider this divergence as a central aspect of Ricardo’s theory.¹⁹ It should be noted that for these scholars it is so whether Ricardo dealt with a monetary system exclusively composed of inconvertible notes (as testified by the title of his famous first monetary pamphlet The High Price of Bullion, A Proof of the Depreciation of Bank Notes) or a system containing (partly or exclusively) convertible notes. By contrast, Benetti and Cartelier boldly assert that “the real distinctive feature of the gold standard is its legal price in terms of £, the consequence of which is that, in a closed economy, the market price of gold cannot diverge from the legal price” (p. 515), and this confirms what they claim in the opening lines of their comments: they accept to bear “the cost” of not being “faithful to Ricardo”.

1.3.1 Can There be a Divergence?

In response to this question, the statement that, in a closed economy (the one, according to Benetti and Cartelier , for which theoretical propositions ought to be demonstrated), the market price of gold is fixed (provided its legal price remains itself fixed) may come as a surprise since in their numerous writings they have (rightly) advocated for long that a theory of prices should not be restricted to equilibrium but account for disequilibrium situations. They maintained in these writings that in a decentralised economy the equilibrium situation cannot be known ex ante by the individual agents, so that a theory of price adjustment is required, which should not violate the decentralised (“market”) character of the economy. In accordance with this motto, they now claim that, to analyse a monetary economy, the non-uniform adjustment of all prices in response to a change in the quantity of money should be made explicit. All except one: the price of gold, which is supposed to be unaffected by such a change.

The reason for this exception is for Benetti and Cartelier crystal-clear: being the standard of money , gold has a legal price, contrary to all other commodities. The market price of gold cannot diverge from the legal price at which gold may be brought to or obtained from the Bank: there cannot be a market for gold because it is impossible that sellers and buyers of it coexist. If the market price of gold were above the legal price, potential buyers would find more advantageous to go to the Bank and they would ignore the market, as potential sellers would do in the symmetrical case of the market price being below the legal price. No individuals face one another about gold: the Bank is the only counterpart. But its relationship with individuals (two-way convertibility ) is not a market one, since the Bank is compelled to give or to accept gold at a legal price.

Leaving aside the consequences of such an assumption for the very notion of standard of money (see below Section “Sraffa’s Theory of Prices and Classical Theory of Money”), it is easy to see why it clashes with Ricardo’s understanding of a monetary economy. Ricardo repeats over and over that the divergence between the market price and the legal price of gold reflects the divergence between the actual quantity of money and that which satisfies “the wants of commerce ”. No individual—and not more the issuing bank itself—may know ex ante the latter. Benetti and Cartelier are right when they say that “the wants of commerce ” depend on what happens in the markets for all other commodities, and it is precisely why—according to their own motto—they cannot be known ex ante. If one assumes (something which is not denied by them) that the Bank can issue more or less liberally the notes demanded by individuals, the decentralised nature of the economy makes it impossible for the Bank to issue an aggregate quantity of notes that would spontaneously equalise with what is required by “the wants of commerce ”. It is thus impossible to reconstruct Ricardo’s monetary theory or to use it as starting point of any Classical theory of money by presupposing, as Benetti and Cartelier do, that “in a closed economy, the market price of gold cannot diverge from the legal price”, if this “closed economy” is a decentralised one.

1.3.2 What Happens in Case of Divergence?

Supposing that the market price of gold is above the legal price (in Ricardo’s words, the currency depreciates), what happens next? Benetti and Cartelier’s answer is: nothing. Nothing between individuals, and I agree. As I wrote in my chapter (on p. 504): “The bank thus provides the missing side of the market for bullion, in the absence of any demand by all other agents who may obtain gold from the bank at the legal price.” But here comes their second critique concerning the market price of gold : nothing will happen between individuals and the Bank either, because this would be assuming that the Bank implements an “odd monetary policy ” (p. 517). In other words, its “economic behaviour is incoherent.” (p. 518) This is a dead blow to a reconstruction that emphasises the coherence of Ricardo’s theory of money.

To see exactly at which level this critique operates, let me again sum up how Ricardo describes the regulation of the issuing of convertible notes by the Bank of England in case of its excess (for details see Deleplace 2017, Chapter 7). The adjustment process is in four steps: (1) Individuals obtain gold for notes at the Bank and export it. (2) If the Bank reissues the notes which are returned to it, there comes a time when its metallic reserve will be viewed by it as too low to guarantee convertibility . It then starts to purchase gold in the market with its notes, in spite of the loss it suffers by buying gold at a higher (market ) price than the (legal) price at which notes are returned to it. (3) Arbitragers stop exporting the gold since it is now more advantageous to sell it to the Bank in the market. (4) This goes on and on (“Penelope effect ”) until the Bank, to stop its losses, contracts its note issue . This contraction is implemented as long as notes are returned to the Bank, that is, as long as the market price of gold is above the legal price. In the end both prices equalise.

Benetti and Cartelier do not object to (1); indeed it is what they retain from Ricardo: “It seems to us that Ricardo clearly favours the regulation of the quantity of money through the export of gold by individuals. Gold is bought by foreigners and the role of the Bank would be to regulate the issuance of money and not to act in the market instead of individuals, as stated by GD.” (p. 518) An ambiguity lies in the word “favours”. If it means that Ricardo contended that money in excess generates an export of gold, no one will deny it; I devoted the whole Chap. 8 of my book to the analysis of how this occurs. If “favours” means that for Ricardo the export of gold should be the preferred means of regulating the note issue , why is it that “the role of the Bank would be to regulate the issuance of money”, as they say? If the export of gold does the job, the sole “role” of the Bank is to give gold passively against its notes when it is required to do so under convertibility . Indeed Ricardo never considered that the Bank should buy and sell gold “instead of individuals, as stated by GD” (I nowhere stated that), but he did consider that the implementation of a specific rule (see below) by the Bank was preferable to the export of gold by individuals as means of regulating the note issue .²⁰

Since the export of gold implies a cost (as for any commodity that must be brought to the market), Benetti and Cartelier cannot either object to (3), that is, the substitution of domestic arbitrage (at no cost, as they suppose) for the international one. Hence their critique does not concern the behaviour of the arbitragers but that of the Bank. They object to (2): “Why does the Bank provide ‘the missing side of the market for bullion’ ([GD] p. 504) whereas this action worsens the problem that motivated its decision to act? [...] The insufficient gold reserve cannot ‘force’ the Bank to apply a monetary policy that reduces it further.” (pp. 517–8) They also object to (4): “Why does not the Bank adopt this measure [the contraction of the note issue ] from the beginning? [...] The demand for gold is positive only if there is at least one agent [the Bank] whose economic behaviour is incoherent.” (p. 518; BC’s emphasis)

Benetti and Cartelier’s critique is entirely right. It just misses its target: their “whys” do not question my reconstruction of Ricardo’s argument but they are in fact addressed by Ricardo himself to the Bank of England . According to him, the Bank behaved stupidly, from its own point of view, as testified by the fact that it was eventually obliged to adopt the opposite behaviour (contracting the issue rather than expanding it to purchase gold).²¹ But mistaking the real target they fail to see that, although the Bank behaves stupidly in the first place, this does not prevent the regulation of the note issue from operating (that is, the divergence between the market and the legal prices of gold will sooner or later disappear, contrary to what happens under inconvertibility), precisely because the “Penelope effect ” obliges the Bank to stop behaving stupidly. In other words, the “incoherent behaviour” of the Bank is indeed “not acceptable”, as Benetti and Cartelier write (p. 518). But it is so, not as a condition of my reconstruction of Ricardo’s argument (which logically would condemn this reconstruction) but, according to Ricardo himself, as the explanation of the poor functioning of the monetary system (poor because the delay with which the Bank changes its behaviour has adverse effects on the economy).

Benetti and Cartelier are also right when they ask, as quoted above: “Why does not the Bank adopt this measure [the contraction of the note issue ] from the beginning?” This is exactly Ricardo’s point, and there is no contradiction between my reconstruction of what happens when it is not so and the contention that all would be better if it were so. On the contrary, this reconstruction justifies Ricardo’s advocacy of his monetary policy rule: the varying of the note issue inversely with the sign of the spread between the market price of gold and its legal price. However, Benetti and Cartelier also fail to see that such adoption, on the one hand ruled out, not only domestic arbitrage (the “Penelope effect ”) but international arbitrage as well (because, in the case of excess note issue , its contraction would occur before the export of gold became profitable), and on the other hand it would dispense the Bank from bothering about the ratio between its metallic reserve and the stock of notes in circulation (because, although returning one’s notes to the Bank was always possible, no one would find advantageous to do it). Ricardo was very clear on these two aspects in his Proposals for an Economical and Secure Currency: “The most perfect liberty should be given, at the same time [as the proposed note convertibility into bullion], to export or import every description of bullion. These transactions in bullion would be very few in number, if the Bank regulated their loans and issues of paper by the criterion which I have so often mentioned, namely, the price of standard bullion, without attending to the absolute quantity of paper in circulation.” (Ricardo 1816, p. 67; reprinted in the 2nd and 3rd editions of Principles, Ricardo 1821, pp. 357–8; emphasis added) Ricardo could stress that he had “so often mentioned” this issuing rule: in the 4th edition of High Price he had already stated: “An unfavourable exchange can ultimately be corrected only by an exportation of goods, by the transmission of bullion,—or by a reduction in the amount of the paper circulation.” (Ricardo 1810–1811, p. 125; Ricardo’s emphasis).²²

At the end of the day, my disagreement with Benetti and Cartelier may be summed-up as follows: one thing is to assume that under convertibility the market price of gold is necessarily equal to its legal price (their position), another thing is to account for the adjustment process through which the market price of gold equalises with its legal price. According to me, this was the specific object of Ricardo’s theory of money, which can thus be called a theory of monetary disequilibrium (meaning by that a situation in which the quantity of money does not “conform” to “the wants of commerce ”).²³

1.4 The Necessary Exteriority of the Standard

Benetti and Cartelier also criticize my proposition that the standard (e.g. gold) is necessarily exterior to the system of production that determines the natural prices of the commodities in the economy where the standard-based currency circulates. According to them, this proposition is both unwarranted and inappropriate.

Quoting my book, they rightly observe that this exteriority of the standard is supposed to result from the contradiction between a determination of the market price of gold by its natural price (if gold belongs to the system of production) and its regulation by a legal price (when gold is the standard of money ). However, referring to their own model (in Benetti and Cartelier 2020), they see no such contradiction so that my proposition is unwarranted. It does not come as a surprise that this contradiction does not show up in their model: as testified by their equation (3*) on p. 520, the money price of any commodity other than gold is equal to the price of this commodity in terms of gold (solution of the Sraffa system in which the numéraire chosen is gold) multiplied by the money price of gold, which they suppose constant since it is the reciprocal of “the quantity of gold which defines the £” (p. 520).

As emphasised in the previous section, Benetti and Cartelier’s assumption that the market price of gold cannot be but equal to its legal price is at odds with Ricardo’s view that it may diverge from it, and that this divergence is caused by the discrepancy between the actual quantity of money and the quantity required by “the wants of commerce ”. Substituting the definitional equality of the market price of gold to its legal price for the regulation of the former by the latter, Benetti and Cartelier boldly eliminate one of the terms of the contradiction, hence the contradiction itself.

After having magically made my exteriority proposition disappear, they set themselves to the task of showing that it is inappropriate for the making of a Classical theory of money consistent with the Sraffa system of natural prices . It is so because of my “neglect of the general problem of gravitation of market prices around natural prices . His [GD’s] focus is on the specific question of the gold market only.” (p. 521; BC’s emphasis) So doing I cannot, however, get rid of this gravitation problem, which hides itself in the quantitative determination of “the wants of commerce ” and in the adjustment of the market price of gold in the foreign mines (“Why should the gold market price be equal to its natural price if no gravitation mechanism is considered?”; p. 522).

Now Benetti and Cartelier do not like gravitation as a theory of the determination of market prices , and they prefer the “Cantillon rule ” according to which the market price of a commodity is determined by the ratio of the quantity of money spent on its market to the quantity of the commodity present on it and entirely sold at that price. This is a position which merits respect but again it misses its target. It is true that “the wants of commerce ” must be expressed at market prices , but I never claimed that my reconstruction of Ricardo’s theory of money accounted for them: my purpose was to elucidate what, following Ricardo, happens when the actual quantity of money diverges from the quantity required by “the wants of commerce ” (unknown by the agents and by the Bank), not to determine the latter, which is to be considered as an independent variable. In the terms used in the previous section, I contend that Ricardo’s theory of money is concerned with monetary disequilibrium (when the actual quantity of money diverges from the level required by “the wants of commerce ”), not with market disequilibrium (when the market prices of commodities diverge from their natural level).

As for the equalization of the foreign market price of gold with its natural price , it may be amusing to note that, in the model I propose of the adjustment in foreign mines (see Deleplace 2017, Chapter 5), the specificity of gold as a commodity demanded for monetary purposes—namely that its demand expressed in money terms should be considered as given, independently of the adjustment process—explains that its market price is determined by what amounts to a “Cantillon rule ” (see equations (5.3) to (5.5) in Deleplace 2017, p. 184). In the gold-producing country as in the gold-importing one, the adjustment of the market price of gold is entirely the consequence of its role in a standard-based monetary economy.²⁴

1.5 Sraffa’s Theory of Prices and Classical Theory of Money

Benetti and Cartelier’s comments—substantiated by their 2020 paper—and my reply suggest that the issue of linking Sraffa’s theory of prices with a Classical theory of a standard-based money may be explored along two alternative routes.

The one I have proposed thanks to a reconstruction of Ricardo’s theory of money focuses on the relationship between the issuing of money and the market price of the standard (gold), which diverges from its legal price when the quantity of money issued is not at the level required by “the wants of commerce ”. Sraffa’s theory of (natural ) prices may then apply at two necessarily distinct levels: in the country producing the commodity used abroad as standard of money (but not in this country) and in the country importing this commodity to use it as standard of money (but which does not produce it). At both levels, a general theory of market prices is unnecessary—this is not the object of the theory of money—and, although the standard of money is a commodity subject (in the gold-producing country) to the Sraffa system, it is outside this system in the country where the currency is based on it.

The other road, suggested by Benetti and Cartelier , assumes that the market price of the standard cannot be but equal to its legal price—in that, although Classical, this road moves away from Ricardo—and focuses on the relationship between the issuing of money and the market prices of all commodities except the one chosen as standard (gold), thanks to the adoption of the “Cantillon rule ”. Sraffa’s theory of (natural ) prices may apply to all commodities produced in a supposedly closed economy, whether the commodity chosen as standard or all other commodities. The theory of money cannot be separated from the theory of (monetary) market prices .

As I see them, the critiques addressed by Benetti and Cartelier to the first road are not conclusive. By contrast, the two main points they contend—the necessary equality between the market and the legal prices of the standard of money and the assumption that gold is produced in the economy which uses it as standard—raise two serious difficulties along the road they advocate. On the one hand, stating that its market price is fixed independently of the quantity of money issued amounts to considering gold as a simple numéraire , and not as a standard regulating the monetary system. Contrary to the title of their 2020 paper, Benetti and Cartelier’s object is not to build a Classical theory of money in which the quantity of money is regulated by a standard but a theory in which it is regulated through the adjustment of the money prices of all commodities except the standard (reduced to a numéraire ). On the other hand, assuming that the “Cantillon rule ” applies to all commodities except the one chosen as standard of money (because its price is fixed) questions the statement that this standard does belong to a system of production of commodities subject to this rule.

More generally, my feeling is that Benetti and Cartelier’s exercise is flawed by the fact that they are trying to pin a Classical-but-not-Ricardian theory of (market ) price adjustment in response to a change in the quantity of money on a Ricardo-Sraffa theory of natural prices . As testified by many others of their writings, published jointly or separately, what they really believe in general is that a Classical theory of prices should not be built on a Ricardo-Sraffa basis but rather along Torrens’s lines, in which the use of the surplus does matter to determine prices (contrary to what happens in Production of Commodities by Means of Commodities). The “Cantillon rule ” that determines market prices is consistent with this view. But in their comments as in their 2020 paper, the exercise is based on Sraffa prices and there lies the root of their ambiguity as regards Ricardo. The question is therefore: can a non-Ricardian Classical theory of money reconcile the “Cantillon rule ” of determination of market prices —which belongs to a non-Sraffian natural habitat—and Sraffa prices?