Exchange liquidity and redemption liquidity

Abstract

Disagreements over the nature of money and consequent confusions regarding liquidity contribute to difficulties integrating monetary theory into the theory of value. For example, an abundance of market liquidity is assumed in asset pricing, whereas a scarcity of monetary liquidity is deemed necessary for consumer price-level determinacy. This paper builds on the insights gained from the evolution of finance to introduce a distinction between exchange liquidity and redemption liquidity as a means of resolving this conceptual dissonance. Both exchange and redemption liquidity can be conceptualised as types of financial option differing in the exercise mechanism offered to the option holder by the option-writer.

1. Introduction

The economics and finance literature displays conceptual dissonance in its treatment of both liquidity and money. An abundance of market liquidity is needed for financial arbitrage and asset pricing, but scarcity of monetary liquidity is needed for consumer price-level determinacy in macroeconomics. Much of this confusion is mirrored in conflicts concerning the nature of money, for which there are broadly two views (Ingham, 2004, 2018; Lawson, 2016). The first, the commodity view, is that the value of money is derived from its metallic or commodity content. The second, the credit view, is that money is essentially tradable credit and is, even in the form of coinage, merely a token that represents a debt payable. Since commodity money is restricted by the physical stock of the commodity, whereas tokens have no recognisable production function, the commodity view can be taken to express ‘the liquidity scarcity view of economics’, whereas the credit view aligns with ‘the liquidity abundance view of finance’ (Mehrling, 2000b, p. 15).

These views of liquidity lead to a ‘fundamental source of misunderstanding’ (Mehrling, 2000b, p. 15) affecting the analysis of liquidity, as is evident in its treatment in both neo-Wicksellian and neo-Walrasian theory. Liquidity scarcity is expressed by the Wicksellian notion that market interest rates must adjust to agree with a natural rate of interest established in the real economy. Thus,

…the persistent attraction to economists of the quantity theory of money, even given all its faults, is that it expresses succinctly the economist’s intuitive sense that the real liquidity of the economy as a whole is scarce and that attempts to increase liquidity by expanding nominal money must eventually reckon with this fundamental real scarcity. (Mehrling, 2000b, p. 15)

On the other hand, the neo-Walrasian ‘theory of value abstracts from the scarcity of liquidity and treats all commodities as equally and perfectly liquid’ (Mehrling, 2000b, p. 15). Consequently,

…monetary theory cannot be brought under the theory of value (as currently constructed) but no alternative analytical structure has yet emerged to gain general acceptance. Until it does, the old distinction between money and credit remains as the economist’s crude theoretical attempt to grapple with the apparent hierarchy of liquidity, with credit viewed as a mechanism for stretching scarce liquidity (money). Similarly, the old attempt to measure the quantity of money remains as the economist’s crude empirical attempt to grapple with the same hierarchy by measuring the size of its base. (Mehrling, 2000b, p. 15)

Both the ‘economics view’ and the ‘finance view’ (Mehrling, 2013b, p. 356), in their attempts to understand the world, ‘resolutely abstract from money’ (Mehrling, 2013b, p. 357) in forming the perfect-exchange constructions of neo-Walrasian general equilibrium (‘trade in goods’) and the perfect liquidity in modern asset pricing (‘trade in financial assets’). Ignoring the monetary or liquidity dimension and ‘the standard practice of focussing attention on the “equilibrium” or “fundamental” exchange rate amounts to focussing on a special limiting case that would prevail if matched-book and speculative dealers were willing to do their work for free’ (Mehrling, 2013b, p. 358).

In attempting to resolve the conceptual dissonance in the treatment of liquidity and restore market makers to a central position in the analysis, this paper establishes the distinction between exchange liquidity and redemption liquidity. This dichotomy can be visualised as two forms of financial option, categorised according to the nature of the liquidity supplied by dealers or financial option writers. The ontological question concerning the nature of money, however, will remain unanswered. Instead, the paper will focus on ‘the current world of credit money’ (Moore, 1988, p. 3) to investigate one of credit-money’s essential features. The reader should then consider any reference to ‘money’ hereafter as referring to credit-money only.

The distinction between exchange and redemption liquidity is outlined in Section 2. Redemption liquidity and exchange liquidity are more fully analysed in Sections 3 and 4, respectively. Section 5 concludes.

2. Liquidity as a financial option

Exchange liquidity and redemption liquidity can be conceptualised as financial options provided by their issuers acting as liquidity providers. By ‘financial option’, it is meant a contract giving the holder the right, but not the obligation, to either purchase (a ‘call’ option) or sell (a ‘put’ option) an asset at a pre-determined (‘exercise’ or ‘strike’) price from the option issuer or ‘writer’. The difference between exchange and redemption liquidity stems from the relationship between the option’s underlying asset and the option’s issuer. The crucial distinction is whether or not the option is exercised over the option-writer’s own liabilities. Exchange liquidity is an option offered over physical or financial assets (such as the liabilities of other agents). By contrast, redemption liquidity is akin to a put option provided by agents over their own liabilities. For example, money issuers, such as banks, offer redemption liquidity by allowing for immediate balance-sheet clearing and reflux at the discretion of the money holder. Other, more recent but perhaps less obvious, examples of assets with redemption liquidity are Exchange Traded Funds (ETFs) and so-called ‘stable coins’.

Exchange liquidity in an organised market can be conceptualised as a combination of call and put options written by a market maker. These options manifest themselves through the public’s entitlement to sell or buy at the ‘exercise prices’ stipulated by the market maker’s respective bid or ask prices. The option premiums embedded in the bid-ask spread compensate the market maker for the provision of these liquidity options. The provider of exchange liquidity has no direct relationship with the asset, but merely facilitates its shiftability, and, as such, exchange liquidity is not an intrinsic property of an asset.

Money, by comparison, is a social relation that has a debt counterpart which exposes it to the potential operation of the reflux mechanism ‘through which excess money is extinguished’ (Lavoie, 1999, p. 104). The redemption liquidity of money arises from its ability to achieve immediate balance-sheet clearing between the issuer of the money and the holder of the money. Money, being a financial instrument arising through the exchange of debts, enables credit clearing and provides the ready means to dissolve its associated social relationship.

The redemption liquidity associated with money contrasts with the exchange liquidity issued by a market maker. Money’s liquidity, or immediate reflux to the issuer, is an essential feature. The option contract underpinning redemption liquidity is exercised at par nominal. And, unlike exchange liquidity, its exercise consists of a mutual debt cancellation, not a physical payment. Yet just like exchange liquidity, redemption liquidity is not an intrinsic property—it must be provided by its issuer. Indeed, instruments with redemption liquidity are not simply limited to the forms of outside money; the placement of state money in a hierarchy of credit allows for a more in-depth analysis of liquidity.

Importantly, the difference between money and non-monetary debt can be delineated by this reasoning: only a credit that can be used to repay an issuer immediately can operate as money. If the issuer is a large economic agent, to which many other agents are constantly indebted, like the state or the banking system, then these credits will be in continuous demand as a means of debt repayment. Although state fiat money appears to be irredeemable, it is sufficient that it can be used to settle debts arising from taxation or to unwind central bank repo operations for it to have redemption liquidity. Similarly, a bank loan can be repaid by means of an offsetting deposit with that same bank, with the result that bank deposits provide redemption liquidity by means of reflux and loan repayment. Both forms of money offer redemption liquidity.

The distinction between exchange and redemption liquidity differs fundamentally from that between market and funding liquidity (Brunnermeier and Pedersen, 2009). Market liquidity describes asset saleability. Funding liquidity refers to the ability to refinance debts as they mature, which is equivalent to selling money forward. Therefore, exchange liquidity encompasses both market and funding liquidity and is microeconomic in nature, since it is an unreliable mirage that disappears in a crisis. Similarly, the distinction between public and private liquidity (Holmström and Tirole, 1998) collapses into two forms of redemption liquidity, by referring to government-supplied and privately supplied pure liquid assets respectively.

Witness the shadow banking system’s failed attempt to create redeemably liquid instruments. The shadow banking system combined market liquidity and funding liquidity to create near-money instruments, in an attempt to increase the supply of redemption liquidity. The innovation of increasing the liquidity supply by means of security repurchase agreements (‘repo’) relied on the ability of exchange liquidity to simulate the redemption liquidity of a cash-equivalent asset. The crucial difference between credit-money and shadow-bank ‘money’ (Gabor and Vestergaard, 2016) is that the liquidity of repo is founded on the market liquidity of its underlying collateral and is therefore reliant on exchange liquidity. The ability to synthesise redemption liquidity from forms of exchange liquidity was tested and subsequently failed during the Great Financial Crisis (Mehrling, 2011).

Overall, the supply of liquidity is offered by dealers that issue liquidity options, and these issuers form a hierarchy (Culham, 2020b). In order of strength, they are generally the government, the central bank, commercial banks and securities dealers, although this order is not immutable. The hierarchy aligns with a public–private (Holmström and Tirole, 1998), outside–inside (Bolton et al., 2011) spectrum of liquidity, and straddles the exchange/redemption dichotomy. The value of each of these options depends on the perception of the default risk of the issuer, determined by the balance of credits payable and receivable (along with any regulatory supervision or explicit liquidity support provided by issuers higher in the hierarchy). For instance, the exchange liquidity offered by a security dealer is unreliable, since it is contingent and can be withdrawn during periods of market turbulence. Nonetheless, all liquidity is provided by a dealer that stands ready to convert assets, and this set of dealers should include traditional money issuers, such as central banks and commercial banks.

3. Redemption liquidity

Money, as it will be treated here, is a credit that can be used to repay a debt to its issuer. It is a financial instrument redeemable on demand with, in effect, a zero time to maturity (Tymoigne, 2014). Or, equivalently, a perpetual asset with a redemption American-style put option written by its issuer. ‘Perpetual’, in this case, means without a pre-defined contractual end of life (Hull, 2003, p. 709), because the holder need never exercise the redemption option, in which case the instrument is everlasting. The effective maturity of a financial instrument able to serve as credit-money is crucial; debts can only be repaid with immediately redeemable or undated credit (Innes, 1913, 1914; Tymoigne, 2017). A debtor can repay a creditor with an offsetting credit against that same creditor, but only if the offsetting credit is currently due for repayment. The creditor can reject as payment their own debts if those debts themselves are not due for payment.

So, if a credit instrument has a maturity date, then it cannot be money since its ability for use as repayment has been postponed or requires a discount to par. For example, demand deposits are money because they can be used to extinguish a loan obligation to a bank. Government bonds, on the other hand, are not money because they must be converted, by means of exchange liquidity, into central bank reserves or state money to be used to pay taxation. Hence central bank reserves, even when they pay interest, differ from bonds, which do not have redemption liquidity.

The price or value of redeemable credits relies on the balance between the amount issued and the availability of offsetting credits in the form of credits returning to the issuer. If more credits are issued than can be supported by available offsets, their value will depreciate (Innes, 1914). By price, it is useful to keep in mind four prices of money as outlined by Mehrling (2018): (1) par, such as the price of converting deposits into currency; (2) interest, the price of future money; (3) exchange-rate, the price of foreign money; and (4) price-level, the price of commodities. Each price is generally associated with a particular type of market maker (Mehrling, 2013a), although an imbalance can affect one or more at once, depending on circumstances.

In the case of state money, for example, this reduction would be due to an imbalance between money and taxation. Unless offsetting credits can be sourced, a debtor (or credit issuer) must redress any credit-debit imbalance by reducing the amount of their debts immediately redeemable. In other words, they will need to retract some of their redeemable credits and replace them with fixed-term debt instruments. This action is a form of own-funding liquidity, whereby own-issued money is sourced. Own-funding liquidity, achieved by selling a new debt instrument, is then subject to the cost of postponing immediate reflux. Pushing repayment dates for debts into the future is an effective method of maintaining the balance of debits and credits currently due. In this way, money issuers can control the reflux mechanism by managing the maturity profile of their debts.

Money can only be refluxed back to its issuer, and the issuer determines how much of its money is available for reflux by liability management through interest rates, thereby maintaining its value to other agents for use as a store of value or a medium of exchange. For an issuer of debt, postponement is the alternative to a reduction in the value of immediately payable debts. The amount of immediately payable debt can be altered by extending it into the future, but at a cost. This cost, expressed as the rate of interest, provides the inducement for some agents to postpone their attempts at reflux. Thus, the issuing of debt, or accessing own-funding liquidity, is itself a form of reflux, necessary to sustain the other prices of money.

This process can be seen in the management of government debt, where swapping state money for non-monetary interest-bearing liabilities removes the need to alter the level of taxation. Government bonds cannot be ‘spent’ because the government has purposely converted its immediate, monetary debts into those not due until a future time and is under no obligation to accept these future debts as payment for current taxes.¹ The value of the government’s immediately redeemable credits, in the form of state money, is based on its current period spending and tax flows.

Similarly, banks, via liability management, carefully maintain their asset-liability term profile so that longer-term assets are funded by longer-term liabilities. Banks cannot create unlimited amounts of monetary liabilities; they require longer-term funding which cannot be used for repaying bank debts or conversion into state money. Banks pay interest on term deposits that cannot be redeemed to prevent the reflux of their liabilities and thereby maintain their value at par with state money.

Money issuers must also be cognisant of their place in the hierarchy of money (Innes, 1914; Bell, 2001; Mehrling, 2013a), which is constructed explicitly in terms of a gold standard or commercial bank convertibility to state money, or implicitly, in terms of an inflation or exchange-rate target. When the nature of the liquidity option offered by money issuers depends on convertibility to higher forms of money, then the resulting hierarchy of money signifies that, unlike exchange liquidity, redemption liquidity is ordinal. The lever for managing the value of money is the interest rate on the non-money debt that the issuer offers to discourage reflux or, equivalently, to prevent the liquidity option from being exercised. Money issuers who aim at a target real value of their money must offer a reflux exit point so that money holders can adjust their desired stock of money. The relevant insight expressed in the endogenous money literature is that, if a money issuer sets the terms under which money-debt can be exchanged for non-money debt, then the balance between money and debt will be determined, not by the issuer, but by the holders.

Similarly, the government, as one of the issuers in the hierarchy of money, must postpone the process of reflux for its monetary liabilities to maintain price-level stability. This mechanism should not be confused with a quantity theoretic explanation of prices. A monetary balance must be maintained primarily against the level of taxation, and not between the stock of money and the productive capacity of the economy, as per the quantity theory of money. At issue is not the size of government debt, or the monetary base, but the liquidity mismatch between taxation and monetary liabilities. These are either resolved by conversion of money into debts or a discount on government liabilities.

Hence, the stock of money comprises immediately payable, or undated, debt, which must be repaid on demand, and so can itself be used for repayment. This stock is not in continuous existence throughout time; it is either repaid, and thereby destroyed, or rolled forward, where rolling a debt forward is fundamentally equivalent to repaying and simultaneously re-borrowing. Conceptually, there is no externally set amount of money that forever circulates with a velocity, instead, there is a continuous flow of creations and cancellations via the reflux mechanism whereby each obligor’s creations and cancellations must balance if their credits are to maintain their value. This value can be either in terms of the unit of account or a real-world benchmark depending on the policy and constraints of the issuer.

Before it refluxes through repayment or replacement with term debt, credit-money emitted by an agent, such as a bank, will circulate among holders. An increase in liquidity preference will increase the demand for credit-money and its circulation time as credits are hoarded and not refluxed. As a consequence, some agents, who are indebted to the credit-money issuer, will struggle to obtain the credits necessary for repayment, and will need to borrow them from others. This scramble for credits will increase the interest rate on the inter-agent borrowing and lending of these credits. Similarly, the cost to the issuer of postponing reflux, or own-funding liquidity, will increase, since the issuer will be competing for credits from these very same hoarders. In response, an issuer high enough in the credit hierarchy can alleviate this shortage, and mitigate the increase in interest rates, by purchasing the financial assets of others with newly created credit-money, as in the case of a central bank conducting daily open-market operations, or quantitative easing during a banking crisis.

The central bank’s convention of using open-market operations or yield-curve control as its monetary mechanism applies a form of liquidity option to those assets that the central bank deals in, mostly government debt. Again, this viewpoint shows us that, government debt is not redeemably liquid; government debt gains exchange liquidity from the central bank’s dealing in the government debt market. The public–private form of liquidity, where government bonds underlie private repo, is an extension of this central-bank provided exchange liquidity (Pozsar, 2014, 2015), as is central support provided to primary dealers during government bond auctions.

4. Exchange liquidity

Market makers and dealers provide exchange liquidity, which is the mechanism by which any pre-established maturity can be reduced for the holder of a financial or physical asset. By providing exchange services for financial assets, for example, dealers mitigate the need to wait until the asset’s maturity for it to become liquid. The dealer can buy the asset outright or provide funding liquidity in the form of a collateralised loan, such as repo. In this way, dealers offer the ability for an agent to adjust the effective maturity of their financial assets.

Exchange liquidity, therefore, encompasses the idea of liquidity when defined as the conversion into money within the three dimensions of certainty, short notice and without loss (Keynes, 1930, p. 67; Hicks, 1989, p. 64). By invoking an idea of loss, this definition of liquidity builds on a theoretical platform of fundamental or long-period prices provided by a general equilibrium model. In practice, however, exchange liquidity often relies on the existence of dealers, who are prepared to provide shiftability based on relative value. These market makers do not attempt to measure fundamental value, nor do their profits rely on assessing the difference between the market price and fundamental value (Mehrling, 2013b, p. 356). Rather, they attempt to operate a ‘matched book’ so that systematic risk is reduced as far as possible (Stigum, 1990, p. 433). It is precisely because exchange liquidity cannot be an intrinsic property of an asset, that analysing liquidity using the three dimensions of certainty, short notice and without loss, which rely on the existence of fundamental value, is paradoxical and inappropriate. These three dimensions also demonstrate that exchange liquidity is ordinal only via a complex multi-dimensional functional relationship. Any ranking of assets in terms of exchange liquidity would be conditional on the dealer structure prevailing at the time of measurement and would therefore only be transitory.

Nor can the concept of exchange liquidity be used to understand the liquidity of cash flows, which, by their very nature, are self-liquidating. A distinction must be made, therefore, between the variability of expected returns across short- and long-dated assets, and any liquidity premium required for an asset supplied with a dealer option. Short-dated assets are more liquid in the sense that their closer maturity date entails an earlier repayment or liquidation date, which a long-dated asset does not. Longer-maturity assets rely on exchange liquidity provided by a dealer in an organised market to be immediately realised in money terms. Exchange liquidity is fundamentally uncertain and corresponds with the empirical observations of time-varying discount factors (Shiller, 1981; Cochrane, 1992). Agents attempt to counter this uncertainty with judicious allocations of assets with price-protection, which provide price stability in the unit of account (Ricks, 2011, 2016), and redemption liquidity.

The deficiencies of analysing liquidity from within a pure exchange or barter model become more apparent when developments in modern finance are considered. These developments enable the layering of financial instruments (such as currency, interest rate, and credit default swaps) over physical assets to disassemble the various risks embedded in the underlying asset (Black, 1970). Consequently, maintaining a focus on assets per se, and not their risks and cash flows, is to suffer from a kind of ‘asset-fetishism’ (Lozano, 2015).

Instead, it is helpful to introduce a distinction between the analysis of ‘classical exchange’ concerning ‘exchange and immediate settlement of physical assets’ (Lozano, 2015, p. 32) and the broader concepts of ‘generic exchange’ and ‘synthetic exchange’ in which settlement is delayed or entirely disconnected from the moment of exchange. Generic exchange produces financial instruments that are ‘non-physical economic objects, whose quality and character of physicality will generally not be informative of its value, for example, loans, bonds, bills and stock’ (Lozano, 2015, p. 33, original emphasis). Their cash flows are decoupled from any reference to physical objects in such a way that the ‘time-horizon of the tenure of exchange is loosened beyond immediate settlement’ (Lozano, 2015, p. 33).

Synthetic exchange broadens this ontological set further by considering risk-transfer instruments such as interest rate and credit default swaps—instruments that have no ‘endogenous limitation’ to their creation (Lozano, 2015, p. 38). In the case of synthetic financial instruments, it is the ‘process of the exchange itself which constitutes the asset, rather than the exchange constituting the process by which some pre-existing assets are exchanged’ (Lozano, 2015, p. 47, original emphasis). For example, the simultaneous creation of a loan and deposit, whereby the deposit itself is money, reveals that banking is the creation of synthetic financial instruments, in other words, the exchange of IOUs (Mehrling, 2011, p. 72) or a derivative security over central bank money (Mehrling, 2000a, p. 404).

The introduction of credit frees the analysis from the restrictions of classical exchange, whereby a physical money object must be involved in the immediate settlement or payment of the exchange. Instead, the tenure of generic or synthetic financial instruments is not limited to immediate settlement, as it is in classical or barter exchange. Credit allows the loosening of the settlement horizon to become an independent phenomenon:

There is virtually no relationship between the legal transaction in which a buyer agrees to pay a certain amount to a seller, in return for an asset owned by the seller or services to be provided by the seller, and the settlement of the transaction in the form of one or more cash flows. (Black, 1970, p. 3)

This ‘extension of the time-horizon of tenure’ (Lozano, 2015, p. 24), made possible by generic and synthetic finance, is distinct from the Arrow–Debreu model of complete futures contracts with payments settled and cleared at the beginning of time. Once this Walrasian classical exchange restriction is removed, a synthetic universe of future settlement contracts opens, revealing an avenue to the fundamental analysis of credit and liquidity. When exchange can be achieved by credit, the tenure of settlement spans an entirely new temporal spectrum of contracts, denominated in the unit of account, relating to the expected delivery of payment, not the delivery of the commodity.

5. Liquidity preference and the rate of interest

The dichotomy between exchange and redemption liquidity can be understood by reference to generic exchange. Liquidity is the product of this structural addition to the dimension of payments and the dynamic interaction with the tenure of exchange. A generic financial asset only provides a guaranteed liquid value at maturity, and therefore imposes a liquidity restriction on its owner. If the owner subsequently wishes to break this contractual tenure, they must appeal to the asset’s exchange value, and this value is vulnerable to the conditions prevailing at the moment of sale (Kahn, 1954). The acceptable tenure of exchange—longer for buyers and shorter for sellers—depends on the balance of preferences for guaranteed redemption liquidity versus uncertain exchange liquidity. These are behavioural drivers independent of intertemporal consumption plans or production possibility frontiers; they are based on the supply and demand for liquidity.

This payment-time dimension, missing from much of the literature on liquidity and money, is governed by the behavioural characteristics underlying tenure preference and allows for a realistic introduction of the rate of interest. By removing the ‘invariance requirement on tenure’ of immediate settlement, entirely new risks appear, since ‘the properties of maturity and interest rate now differentiate themselves, and the image of the value of the object is free to grow or shrink over the time-horizon of the exchange’ (Lozano, 2015, p. 3). Interest rates, as a truly monetary phenomenon, only appear with generic financial instruments (Lozano, 2015, p. 3 and p. 126) and their basic property of tenure undermines theories of the rate of interest in the tradition of Real Analysis (Schumpeter, 1954, p. 277). The timing of cash flows is independent from the moment of exchange and, therefore, consumption itself, and so it follows that the monetary rate of interest cannot represent a trade-off between consumption in the present and the future.

In models with a single fixed moment of exchange, such as that associated with a Walrasian auctioneer or portfolio-rebalancing models (Tobin, 1969), the concept of a risk-free asset is sufficient to function as the determinant of the rate of interest, since the requirement of immediate settlement is strictly enforced. Without these restrictions on exchange settlement, the rate of interest is required to equilibrate the supply and demand for generic financial instruments. Money, as a credit with an immediate redemption option, can therefore be situated on a credit spectrum of delayed settlement that expands the tenure of exchange. Money is the financial instrument that preserves the requirement of immediate settlement associated with classical exchange, thereby eliminating the risk associated with a decoupled time horizon.

This dimension of credit and liquidity—and the rate of interest that arises with it—is unrelated to assets per se. The cash flows associated with any particular asset can be decoupled from it, and the creation of generic and synthetic instruments, which separate risk from assets, confirms that liquidity is also not a property of any specific asset; it is a property of the nature and the maturity of cash flows. It is informed by the knowledge that the value of generic financial instruments, although perhaps conceptually in equilibrium at any point in time, can change over their life.

The extra time dimension related to generic and synthetic finance creates entirely new risks missing from Real Analysis. For example, liquidity preference is not satisfied by assets, such as long-term Treasury bonds, that trade in, what are often referred to as, deep and liquid markets and can be realised for cash at short notice. Although such an asset can be realised for an amount of cash that is (presumably) close to its ‘fair value’, the actual amount of cash is unpredictable and is therefore not suitable for meeting an unexpected cash payment of possibly knowable size. Keynes (1936) initially presents his theory of liquidity preference as an asset allocation decision in a model comprising only two types of asset: ‘money’ or ‘cash’; and ‘debts’ or ‘bonds’. It is for more than pedagogical reasons that Keynes draws this sharp distinction between liquid assets and non-liquid assets (Culham, 2020a). Liquidity preference is not founded on the market liquidity or funding liquidity displayed by bonds; it expresses the desire for assets with price-protection, ‘information insensitivity’ (Dang et al., 2012), or redemption liquidity. Liquidity preference motivates investors to hold assets that maintain their value against the unit of account, instead of exchange liquidity.

As identified by Bibow (1998), any premium associated with exchange liquidity should be interpreted as a holding cost, with the consequence that the large empirical literature on bid-ask spreads is not directly relevant to the theory of liquidity preference. Analysis of the dealer microstructure provides insights into the profit-making activities and asset inventory management of dealers but does not contribute to a macroeconomic theory of liquidity. Ultimately, the microstructure analysis can only reveal the various ways that the carrying cost is expressed: either as a profit to market makers to provide an exchange-liquidity option; or as the cost of search and delay in transactions.

Instead, liquidity is realised as an abstract premium earned by the issuer of assets with redemption liquidity, known as seigniorage or brassage. Money, being the liquid asset par excellence, is awarded this liquidity premium, not illiquid assets. In reality, however, money earns no premium. The fee from lending money arises from not actually holding money, it arises from holding another type of asset entirely: a loan for which the rate of interest is its return. The failure to distinguish the rate of interest from the liquidity premium of money is a subtle, but significant, error, which has led the interpretation of liquidity preference astray. The liquidity premium that Keynes assigns to money is due to the utility gained by actually holding it (Kregel, 1998); the premium is not directly apparent in its price. The fact that this utility is observed through its effect on the expected return of other assets is an implication of the theory but not the core of the theory itself. The utility attached to the liquidity of money makes it costly to borrow from other agents, and interest rates are evidence of this utility but not its cause. In equilibrium, and abstracting from other risks as Keynes does in the two-asset framework, the rate of interest earned from lending money must equal the liquidity premium on money, but they are not the same thing.

The price-protection feature, in a probabilistic framework, is the property that an asset maintains a constant price across all future states and sample paths of nature [and also to changes in the state of long-term expectation (Hayes, 2018)]. The desire for liquidity is to avoid uncertainty in future returns and the time-variation in discount factors. The phenomenon behind liquidity preference is precisely the unpredictability of the discount factor and asset-price risk. Redemption liquidity provides a stable value in all outcomes, especially in bad states of nature, where the value of other assets has fallen.

Price-protected assets have a payoff that displays zero correlation with systematic risk and a low risk of default, and this information insensitivity increases their predictability and, hence, liquidity. The zero-correlation feature is commonly associated with the concept of the risk-free rate, which, unlike the rate generally earned on money-proper, is non-zero. The risk-free rate, however, is only appropriate for a well-defined investment horizon, without path-dependency. For example, selling a risk-free bond before its maturity relies on exchange liquidity, which exposes the owner to price risk. In the complete-markets model that forms the theoretical foundation for the risk-free rate, exchange liquidity poses no problem; all assets are tradable at their fundamental values. In reality, however, it is the interplay between exchange liquidity and redemption liquidity that determines the rate of interest.

6. Conclusion

In an attempt to resolve the literature’s conceptual dissonance in its treatment of liquidity, this paper has introduced a fundamental distinction between exchange liquidity and redemption liquidity. Each type of liquidity is associated with financial options that offer liquidity services to their owner. Neither exchange liquidity nor redemption liquidity is an essential property of any specific asset, but are differentiated by the mechanism by which the liquidity option is offered to the asset owner.

Exchange liquidity covers the tradability offered by dealers in an organised market who extract profits by mitigating the costs of search and delay in exchange. These profits are the reward to dealers for providing exchange liquidity on the assets and liabilities of others and translate to a user cost for their customers. Redemption liquidity, by contrast, arises in the specific conditions related to reflux and liability management of money issuers.

Once this liquidity dichotomy is identified, the principles underlying the theory of liquidity preference are clarified, and the validity of Keynes’s two-asset model can be recognised. Liquidity preference, in the modern credit-money economy, exists on the cusp between exchange and redemption liquidity and operates to balance the relative supply and demand for each. The balancing factor is the rate of interest and it is here that the difference between theory of liquidity preference, based in the tradition of Monetary Analysis, and the interest rate theories of Real Analysis, becomes most apparent. Real Analysis, where perfect-exchange liquidity is awarded to all assets, cannot encompass or comprehend this dichotomy. The hypothetical Walrasian auctioneer, by clearing all credit at the beginning of time, entirely abstracts from the risks associated with the change in value of delayed cash flows provided as means of payment or investment. This abstraction from the risks associated with the tenure of exchange removes any reason to favour redemption liquidity over exchange liquidity.

The recognition that the tenure of exchange can expand to allow for more than immediate settlement transforms the understanding of both liquidity and money. Delayed settlement—realised in the form of generic financial assets—introduces credit at a fundamental level. Money appears in many forms as part of a dynamic credit hierarchy and its social-relational aspect comes to the fore. The result is a structure and dimension of payments vastly beyond that embraced by models based on pure exchange and commodity utility that take the focus in Real Analysis.

Keynes’s theory of liquidity preference provides the point of departure for a conceptual framework organising these characteristics of liquidity. With the distinction between redemption liquidity and exchange liquidity made explicit, the theory of liquidity preference and Keynes’s two-asset model are sufficient to present a theory of the rate of interest in the tradition of Monetary Analysis whereby the uncertainty regarding the future rate of interest determines the current rate of interest.

Thanks to Jerry Courvisanos, John E. King, Éric Tymoigne, Warwick Smith and the referees for many helpful comments. This argument was first presented in the author’s PhD thesis: Culham, James. 2018. A Conceptual Framework for a Theory of Liquidity, unpublished doctoral thesis, Federation Business School, Federation University Australia. All opinions and errors are my own.

Footnotes

Bibliography