Safe Assets as Commodity Money

• Published date: 01 September 2019
• Author: MAYA EDEN,BENJAMIN S. KAY
• Date: 01 September 2019
• DOI: http://doi.org/10.1111/jmcb.12560

DOI: 10.1111/jmcb.12560

MAYA EDEN

BENJAMIN S. KAY

Safe Assets as Commodity Money

This paper presents a model in which safe assets are systemic because they

are the medium of exchange in risky assets. It connects the literature from

banking and finance on safe assets to the monetary literature on alternative

monetary systems involvingcommodity money, interest bearing money, and

private money creation. Because safe assets haveintrinsic value, changes in

their supply lead to changes in market efficiency.Additionally, because safe

assets are costly to produce, there is overproduction of safe assets relative

to the social optimum. When the model is calibrated to plausible liquidity

premiums the resulting inefficiencies are not large.

JEL codes:E41, E42, G11, G12

Keywords:privatemoney creation, liquidity, institutional money

alternatives.

THE OBJECTIVE OF THIS PAPER is to characterize the efficiency

properties associated with the use of safe assets as a medium of exchange. We model

safe assets as money used by portfolio managers to adjust their portfolio holdings.

As a medium of exchange, safe assets have two important features that distinguish

them from pure fiat money: first, they carry coupon payments. Second, while the

production of fiat money requires only a printing press (if that), the production of

safe assets requires real investment in safe projects, and perhaps also real resources

devoted to monitoring or packaging these projects into tradable assets.

These two features are characteristic of commodity monies. For example, gold can

be thought to carry coupon payments, as it is valued not only for its use as a medium

The views and opinions expressed in this paper are solely the responsibility of the authors and should

not be interpreted as reflecting the official policy or position of World Bank, the Board of Governorsof

the Federal Reserve System, or anyone else associated with the Federal Reserve System. Wethank Greg

Duffee, as well as seminar participants at the WorldBank and Office of Financial Research for their helpful

comments and suggestions. All remaining errors are our own. This research did not receive any specific

grant from funding agencies in the public, commercial, or not-for-profit sectors.

MAYA EDEN is an assistant professor of economics at Brandeis University (E-mail: meden@

brandeis.edu). BENJAMIN S. KAY is a senior economist at the FederalReserve Board (E-mail: Benjamin.S.

Kay@frb.gov).

Received October 27, 2016; and accepted in revised form August 17, 2018.

Journal of Money, Credit and Banking, Vol. 51, No. 6 (September 2019)

Published 2018. This article is a U.S. Government work and is in the public domain in the

USA.

1652 :MONEY,CREDIT AND BANKING

of exchange, but also for its beauty and malleability. Additionally, the production of

gold is costly, as its extraction requires investment in excavation and machinery. The

inefficiencies associated with commodity money (Barro 1979, Kiyotaki and Wright

1989, and Ritter 1995) apply to safe assets as well. Particularly, we show that (i)

unless the coupon payments on safe assets are high enough, there is suboptimal

trading; and (ii) the economy expends too many resources on the production of safe

assets. Thus, in principle, safe assets may be a less-efficient medium of exchange

than a fiat currency.

We calibrate our model to assess the magnitudes of these inefficiencies. Our cal-

ibration suggests that the use of safe assets as a medium of exchange is close to

efficient, in the sense that there is only a small difference in welfare between the

current system and one in which there is an ideal medium of exchange. Given plau-

sible estimates of the liquidity premium, our simulations suggest that risk sharing is

close to efficient and that there are only minuscule inefficiencies generated by the

overproduction of safe assets.

We also use this calibration to study a sharp contraction in safe assets, analogous

in magnitude to what occurred in the 2007–09 crisis.1In our model, a contraction in

the supply of safe assets reduces risk sharing, which has the potential for generating

large welfare losses. However, we find that, given plausible parameterizations, the

welfare losses from this contraction are small. This is because the contraction in

trading volume (in units of assets) is much smaller than the reduction in safe assets.

Counterintuitively,while the spike in liquidity premiums during the crisis was viewed

by many as cause for alarm, our model suggests that this equilibrium adjustment had

an important mitigating effect. This suggests that the collapse in safe assets during

the crisis is unlikely to be the direct cause of its subsequent severity, at least not

through the channels emphasized here.

While there has been broad agreement in the literature that the analogy between

safe assets and money is potentially useful, there has been some disagreement re-

garding the extent to which safe assets should be thought of as “nominal balances”

or as “real balances” (Mand M/prespectively in standard notation). For example,

Stein (2012) and Krishnamurthy and Vissing-Jorgensen (2012) consider money-

in-the-utility-function models in which safe assets enter the utility function directly,

analogously to M/p. In contrast, Rocheteau and Wright (2013), Midrigan and Philip-

pon (2016), Hart and Zingales (2011, 2015) consider models in which safe assets

are the medium of exchange, analogously to nominal balances. Our paper clarifies

the relationship between these two approaches by showing that, as a form of com-

modity money, safe assets share some features with both real and nominal notions of

money. This paper contributes to an emerging literature on the systemic importance

of safe assets, including Caballero (2006), Caballero, Farhi, and Gourinchas (2008),

Gourinchas and Jeanne (2012), Gorton and Ordonez (2013), and Dang, Gorton, and

1.We emphasize that our model is designed to study the equilibrium effects of a contraction in safe

assets on trading volumes and risk sharing. It is ill-equipped to comment on what may have caused such a

contraction; for a discussion of possible mechanisms, see Kacperczyk and Schnabl (2010).

MAYA EDEN AND BENJAMIN S. KAY:1653

Holmstrom (2012), among others. Similar to Rocheteau and Wright (2013), Shen

and Yan (2015), and Hart and Zingales (2011, 2015), this paper contributes to the

discussion by studying the money-like properties of safe assets. Most closely related

are Hart and Zingales (2011, 2015). These papers highlight that, when safe assets

have a transaction role, there is an oversupply of safe assets relative to the social

optimum. In this paper, we apply this insight to an environment in which safe assets

are used for facilitating trading in risky assets.

This setting is similar to the “cash in the market” literature (Allen and Gale

1994, 2005, and Acharya and Yorulmazer 2007). In a “cash in the market” setting,

asset prices depend on the liquidity of market participants. When investors and

intermediaries have lots of cash on hand, liquidity trades have little effect on prices.

When they have very little cash on hand, even small shocks can have large effects on

prices. Similarly, in this paper, when safe assets are plentiful, the liquidity premium

is low, asset prices are high, and trade is plentiful. Conversely, when safe assets are

scarce, the liquidity premium is higher, prices lower, and trade more subdued. A

key difference is that, in this setting, these channels exist even without aggregate

credit risk, with full market participation, and in a setting that can be calibrated

to observable economic and financial market quantities to simulate the economic

and welfare consequences. We also introduce a feature, novel to this literature, of

endogenous but costly creation of safe assets by market participants.

This paper is related to an extensive literature that studies pecuniary externali-

ties in constrained environments, such as Geanakoplos and Polemarchakis (1986),

Greenwald and Stiglitz (1986), Caballero and Krishnamurthy (2001), Lorenzoni

(2008), Bianchi (2011), Farhi, Golosov, and Tsyvinski (2009), Bengui (2013), Ko-

rinek (2011), Eden (2016) and D´

avila and Korinek (2017), among others. It is a

well-established principle that, in the presence of binding constraints, a decentralized

equilibrium may be inefficient due to inefficient price impacts on constrained agents.

This paper relates this principle to the inefficiency of privatemoney creation in an en-

vironment in which the cash-in-advance constraint is binding, and draws implications

regarding excessive private creation of safe assets.

Our modeling approach is motivated by important insights from the New Mone-

tarist view of liquidity (see Lagos, Rocheteau, and Wright 2017 for a review). As

illustrated by Lagos (2011), there is a theoretical equivalence between collateral as-

sets and assets used as a medium of exchange. In our model, we assume that safe

assets are used directly as a medium of exchange, building on this equivalence for

the broader interpretation of the model. While our approach resembles a New Mon-

etarist approach in some ways, there are also some differences. In particular, our

model departs from the assumption of bilateral trading and assumes the presence of

trading posts in which risky assets are exchanged for the safe asset. This precludes

the possibility that traders meet by chance and choose to exchange one risky asset

for another.2

2.This trading structure appears realistic in the context of high-paced financial markets, in which the

number of participants is large relative to the number of assets (though perhaps less so in the context of