A Model of Monetary Policy and Risk Premia

• Date: 01 February 2018
• Published date: 01 February 2018
• DOI: http://doi.org/10.1111/jofi.12539

THE JOURNAL OF FINANCE •VOL. LXXIII, NO. 1 •FEBRUARY 2018

A Model of Monetary Policy and Risk Premia

ITAMAR DRECHSLER, ALEXI SAVOV, and PHILIPP SCHNABL∗

ABSTRACT

We develop a dynamic asset pricing model in which monetary policy affects the risk

premium component of the cost of capital. Risk-tolerant agents (banks) borrow from

risk-averse agents (i.e., take deposits) to fund levered investments. Leverage exposes

banks to funding risk, which they insure by holding liquidity buffers. By changing the

nominal rate the central bank influences the liquidity premium, and hence the cost

of taking leverage. Lower nominal rates make liquidity cheaper and raise leverage,

resulting in lower risk premia and higher asset prices, volatility, investment, and

growth. We analyze forward guidance, a “Greenspan put,” and the yield curve.

IN TEXTBOOK MODELS (E.G., WOODFORD (2003)), monetary policy works by chang-

ing the real interest rate. Yet a growing body of empirical evidence shows that

monetary policy also has a large impact on the risk premium component of

the cost of capital.1Moreover, many central bank interventions can be use-

fully interpreted as targeting risk premia. For instance, a “Greenspan put” in

the 1990s and low interest rates in the mid-2000s arguably led to excessive

leverage and compressed spreads.2During the financial crisis, large-scale as-

set purchases, equity injections, and asset guarantees were all explicitly aimed

at supporting risky asset prices (see Bernanke (2013) for a discussion). Since

the crisis, with spreads near historic lows, an important debate has centered on

whether low interest rates fuel “reaching for yield” and as a result pose a threat

to financial stability (Stein (2014)). These observations point to an underlying

risk premium channel of monetary policy.

∗The authors are from New York University Stern School of Business and NBER. Schnabl is

also with CEPR. We thank Ken Singleton (the Editor), two anonymous referees, Viral Acharya,

Xavier Gabaix, John Geanakoplos, ValentinHaddad, Matteo Maggiori, Alan Moreira, Stefan Nagel,

Francisco Palomino, and Cecilia Parlatore, as well as participants at the 2012 CITE conference,

the 2013 Kellogg Junior Macro conference, the Princeton Finance Seminar, the 2013 Five-Star

Conference at NYU, the 2014 UBC Winter Finance Conference, the 2014 Cowles GE Conference

at Yale, Harvard Business School, the 2014 Woolley Centre conference at LSE, the Minneapolis

Fed, the 2015 AFA Meetings, the 2015 FARFEConference, the New York Federal Reserve, and the

Federal Reserve Board. The authors have read the Disclosure Policy of the Journal of Finance and

have nothing to disclose.

1Bernanke and Kuttner (2005) show that monetary policy surprises have a large impact on

stock prices and that this impact primarily reflects changes in risk premia. Hanson and Stein

(2015) and Gertler and Karadi (2015) find parallel results for long-term bond yields and credit

spreads. Gilchrist and Zakrajˇ

sek (2012) find that changes in risk premia have a strong influence

on the macroeconomy.

2See, for example, Blinder and Reis (2005), Rajan (2011), and Yellen (2011).

DOI: 10.1111/jofi.12539

317

318 The Journal of Finance R

In this paper, we develop a dynamic asset pricing model of the risk premium

channel of monetary policy. In the model, taking leverage exposes financial

institutions to funding shocks that require them to liquidate assets. To avoid

engaging in costly fire sales, they hold buffers of liquid securities that can be

sold rapidly at full value. Consequently, the cost of taking leverage depends on

the cost of holding liquid securities—the liquidity premium. The central bank

governs this liquidity premium by varying the nominal interest rate. A low

nominal rate leads to a low liquidity premium, a relationship that has strong

empirical support. In turn, a low liquidity premium decreases the cost of taking

leverage and hence increases risk-taking, which reduces risk premia and the

cost of capital in the economy.

Our model features an economy populated by two types of agents who differ

in their risk aversion. We think of the more risk-tolerant agents as pooling

their wealth into the net worth (equity capital) of financial institutions, or

banks for short. In equilibrium, banks take levered positions in risky assets by

borrowing from the more risk-averse agents using short-term risk-free claims,

which we think of as taking deposits. Our view of banks as levered risk-takers

is purposely simplified, abstracting from other functions such as screening and

monitoring in order to focus on risk-taking and risk premia. This simplified

view has the advantage of accommodating a diverse set of financial institutions,

including commercial banks, broker-dealers, and hedge funds, whose unifying

characteristic is that they take leverage using short-term debt.

Taking deposits exposes banks to funding (rollover) risk (e.g., Allen and Gale

(1994)). When hit by a funding shock, banks are forced to redeem a fraction

of their deposits. To do so, they must immediately liquidate some of their as-

sets. Liquidating risky assets rapidly is costly because it leads to fire sales.

To avoid this, banks hold buffer stocks of liquid securities, which can be liq-

uidated rapidly at full value. Thus, to insure against losses in the event of a

funding shock, banks set aside a fraction of each deposit dollar they raise and

hold it in liquid securities. In this way, the risk of funding shocks creates a

complementarity between holding liquidity and taking leverage.

We model two types of liquid securities: central bank reserves, which have

the highest level of liquidity, and government bonds. Banks’ demand for liq-

uidity buffers causes liquid securities to command a premium in equilib-

rium. This liquidity premium depends on the nominal interest rate. The

liquidity premium of reserves equals the nominal rate because that is the

opportunity cost of holding them. The liquidity premium of government

bonds is likewise proportional to the nominal rate because government bonds

and reserves are substitutable sources of liquidity. Therefore, by changing

the nominal rate, the central bank changes the cost of holding all liquid

securities.3

3Recently, interest on reserves (IOR) has attracted significant attention. When reserves pay

interest, reserves’ liquidity premium equals the difference between the nominal rate and their

rate of interest. In this case, the central bank targets this difference rather than the full level of

the nominal rate. Section III.B provides further discussion.

A Model of Monetary Policy and Risk Premia 319

0%

5%

10%

15%

20%

25%

30%

35%

40%

-0.5%

0.0%

0.5%

1.0%

1.5%

2.0%

2.5%

3.0%

3.5%

4.0%

4.5%

5.0%

5.5%

1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010

Fed Funds-T-Bill spread Fed Funds rate (right axis)

Figure 1. The Fed funds-T-bill spread and the Fed funds rate. The figure plots the spread

between the fed funds rate and the three-Month T-bill (solid red, left axis), and the fed funds rate

(dashed black, right axis). Monthly data, 1955 to 2010. (Color figure can be viewed at wileyonlineli-

brary.com)

Figure 1examines this prediction empirically. It plots the nominal short

rate, as measured by the Fed funds rate, against the liquidity premium on

government bonds, as measured by the spread between the Fed funds rate

and the three-month T-bill rate, from 1955 to 2010. As the figure shows, the

relationship between these two series, a rate and a spread, is very strong. Their

correlation is 78%, and they exhibit tight comovement both in the cycle and in

the trend, consistent with the transmission of the nominal rate to the liquidity

premium that we model.

The central bank’s ability to influence risk-taking works through this trans-

mission mechanism. When the central bank raises the nominal rate, the higher

liquidity premium increases banks’ cost of taking leverage and hence reduces

their risk-taking. The result is a decrease in the overall demand for risk-taking

in the economy, an increase in the effective aggregate risk aversion, and ulti-

mately an increase in risk premia.

Monetary policy in our model takes the form of a nominal interest rate rule,

which is a function of the single state variable—the share of banks’ net worth

of the total wealth in the economy. We consider a number of interest rate rules

and analyze their positive implications for equilibrium prices and quantities.4

4We do not take a stance on optimality because doing so requires making two strong assump-

tions. The first is the choice of welfare criterion, which is difficult in our incomplete-markets