Consumption, asset wealth, equity premium, term spread, and flight to quality

• Author: Mauro Costantini,Ricardo M. Sousa
• DOI: http://doi.org/10.1111/eufm.12243
• Published date: 01 June 2020
• Date: 01 June 2020

Eur Financ Manag. 2020;26:778–807.wileyonlinelibrary.com/journal/eufm778

|

© 2019 John Wiley & Sons Ltd.

DOI: 10.1111/eufm.12243

ORIGINAL ARTICLE

Consumption, asset wealth, equity premium,

term spread, and flight to quality

Mauro Costantini

1

|

Ricardo M. Sousa

2,3

1

Department of Industrial and

Information Engineering and Economics,

University of L'Aquila, L'Aquila, Italy

2

Department of Economics and

Economic Policies Research Unit (NIPE),

Gualtar Campus, University of Minho,

Braga, Portugal

3

LSE Alumni Association, Houghton

Street, London School of Economics and

Political Science, London, United

Kingdom

Correspondence

Ricardo M. Sousa, University of Minho,

Department of Economics and Economic

Policies Research Unit (NIPE), Campus

of Gualtar, Braga, Portugal; London

School of Economics and Political

Science, LSE Alumni Association,

Houghton Street, London WC2 2AE,

United Kingdom.

Email: rjsousa@eeg.uminho.pt and

rjsousa@alumni.lse.ac.uk

Abstract

We link transitory deviations of consumption from its

equilibrium relationship with aggregate wealth and labor

income to equity returns on the one hand, and to two

characteristics of bond investors—the premium de-

manded to hold long‐term assets, and “flight to quality”

behavior—on the other hand. Using a panel of 10 euro

area countries over the period 1984Q1–2017Q4, we show

that a rise in the consumption–wealth ratio predicts both

higher equity returns and the future term spread, while a

fall in the consumption–wealth ratio explains a large

fraction of the rise in the spread between the “risky”and

the “safe‐haven”bond.

KEYWORDS

asset wealth, consumption, flight to quality, labor income, panel data,

term premium

JEL CLASSIFICATION

C33; E21; E44; D12

EUROPEAN

FINANCIAL MANAGEMENT

We are grateful to the Editor, John Doukas, and an anonymous referee, to participants in the 26th International

Conference on “Forecasting Financial Markets”and the 50th Anniversary Conference of the Money, Macro and

Finance (MMF) Group at the London School of Economics and Political Science (LSE), to seminars and discussions at

Queen Mary University of London and Lille University, and to Deven Bathia, Benoît Demil, Alain Desreumax, Brigitte

Granville, Praveen Gupta, Fredj Jawadi, Eleni Lioliou, Lin Ma, Sushanta Mallick, Pedro Martins, and Gulnur

Muradoglu, for their constructive comments and suggestions that considerably improved this paper. NIPE's work is

financed by National Funds of the FCT—Portuguese Foundation for Science and Technology within project “UID/

ECO/03182/2019”.

1

|

INTRODUCTION

A well‐established body of the finance literature builds the theoretical linkage and explores the

empirical relationship between the dynamics of macroeconomic activity and asset wealth. Some

authors start with a capital asset pricing model (CAPM) and emphasize that the (equity) risk

premium reflects an asset's ability to insure against unfavorable consumption fluctuations

(Lintner, 1965; Sharpe, 1964). Seminal studies also rely on the concept of discounted present

value and derive structural associations between macroeconomic data and future equity returns

(Campbell, 1996; Campbell & Shiller, 1987).

Other authors enrich the standard CAPM model with additional factors that improve asset

pricing accuracy. These include the three‐factor (i.e. market, size, and value risks) model of

Fama and French (1992); the four‐factor models of Carhart (1997) and Pastor and Stambaugh

(2003), which add momentum and liquidity risks, respectively, to the three‐factor model; the

five‐factor model of Fama and French (2015), which includes operating profitability and

investment in the three‐factor model; the six‐factor model of Fama and French (2018), which

adds momentum to the five‐factor model; the fourth‐quarter consumption growth model of

Jagannathan and Wang (2007); and the financial intermediaries' leverage factor model of

Adrian, Etula, and Muir (2014).

Along the same lines, but more recently, new factor pricing models have tried to explain the

cross‐section of expected equity returns. Starting with the inclusion of market and size factors,

the four‐“mispricing”‐factor model of Stambaugh and Yuan (2017) adds management and

performance risks; the

q

‐factor model of Hou, Xue, and Zhang (2015) incorporates the

investment‐to‐assets ratio and the return on equity, and the

q

5

‐factor model of Hou, Mo, Xue,

and Zhang (2019) accounts for expected growth in the

q

‐factor model. Single‐factor models,

such as those by Kuehn, Simutin, and Wang (2017), Wen (2019) and Lettau, Ludvigson, and Ma

(2019), focus on labor search frictions, aggregate asset growth, and capital share risk,

respectively. Finally, the three‐factor model of Daniel, Hirshleifer, and Sun (2019) contains

market, (net share issuance) financing and post‐earnings‐announcement‐draft risks.

Yet, given the CAPM's difficulty in matching differences in the covariance of equity returns

and contemporaneous consumption growth with differences in expected equity returns (Fama

& French, 1992; Hansen & Singleton, 1982),

1

some research highlights the role of long‐term

risks (Bansal & Yaron, 2004; Bansal, Dittmar, & Lundblad, 2005, 2016; Parker & Julliard, 2005):

risk compensation should mirror the exposure of long‐term consumption to the full path of

future cash‐flows and discount rates (Campbell, Polk, & Vuolteenaho, 2010; Campbell &

Vuolteenaho, 2004). These are the pillars of the “intertemporal capital asset pricing model”

(Campbell, 1993).

2

Another research avenue reconciles the joint dynamics of consumption growth and

equity returns by delving into alternative specific preferences of the representative

investor (Campbell & Cochrane, 1999; Constantinides, 1990; Pakos, 2011; Piazzesi,

Schneider, & Tuzel, 2007; Yogo, 2006). These studies lie at the heart of “consumption

1

As noted by Antoniou, Doukas, and Subrahmanyam (2016), such inability could reflect the stronger presence of overconfident and unsophisticated traders

during optimistic periods and less noise trading during pessimistic periods.

2

For a critical assessment of the CAPM model, see Levy (2010) and Zhang (2017). Subrahmanyam (2010) evaluates crucial aspects of stock return predictability,

including correlational structure and heterogeneity of controls and methodology. Barillas and Shanken (2018) compute Bayesian probabilities for all asset

pricing models based on subsets of a given number of candidate traded factors. They show that models including momentum, value, and profitability factors

tend to dominate.

COSTANTINI AND SOUSA EUROPEAN

FINANCIAL MANAGEMENT

|

779