Financial Markets, the Real Economy, and Self‐Fulfilling Uncertainties

• Author: XUEWEN LIU,PENGFEI WANG,JESS BENHABIB
• DOI: http://doi.org/10.1111/jofi.12764
• Published date: 01 June 2019
• Date: 01 June 2019

THE JOURNAL OF FINANCE •VOL. LXXIV, NO. 3 •JUNE 2019

Financial Markets, the Real Economy, and

Self-Fulfilling Uncertainties

JESS BENHABIB, XUEWEN LIU, and PENGFEI WANG∗

ABSTRACT

We develop a model of informational interdependence between financial markets

and the real economy, linking economic uncertainty to information production and

aggregate economic activities in general equilibrium. The mutual learning between

financial markets and the real economy creates a strategic complementarity in their

information production, leading to self-fulfilling surges in economic uncertainties.

In a dynamic setting, our model characterizes self-fulfilling uncertainty traps with

two steady-state equilibria and a two-stage economic crisis in transitional dynamics.

UNCERTAINTY IN BOTH FINANCIAL MARKETS and the real economy rises sharply

during recessions. The recent financial crisis of 2007 to 2009 presented one of

the most striking episodes of such heightened uncertainty. Financial market

uncertainty, measured by the VIX index, jumped by an astonishing 313% in

the Great Recession. The increase in measured real economic uncertainty was

equally astounding. For instance, macroeconomic uncertainty measured in Ju-

rado, Ludvigson, and Ng (2015) almost doubled, and Bloom et al. (2012) report

a 152% increase in micro-level real uncertainty as measured by the firm-level

dispersion of output. What causes such sudden spikes in uncertainty? Why do

financial uncertainty and real uncertainty move together? And why do they

rise sharply in recessions? These questions are of central importance to under-

standing the interaction between financial markets and the real economy. In

this paper, we provide a theoretical framework for addressing these questions.

Specifically, we develop a model of informational interdependence between

financial markets and the real economy, linking uncertainty to information

∗Jess Benhabib is in the Department of Economics, New York University. Xuewen Liu is in the

Department of Finance, Hong Kong University of Science and Technology. Pengfei Wang is in the

Department of Economics, Hong Kong University of Science and Technology. We are grateful to

the Editor and two referees for their detailed comments and suggestions that have significantly

improved the paper. We thank Russell Cooper; Feng Dong; Gaetano Gaballo; Shiyang Huang;

Alexander Kohlhas; Wei Xiong; Liyan Yang; Kathy Yuan; Zhen Zhou; and seminar participants

at the FIRS conference 2018, Wharton Conference on Liquidity and Financial Fragility 2017,

Barcelona GSE Summer Forum 2017: Information frictions, HKUST Macro Workshop 2017, Ts-

inghua Finance Workshop 2017, AFR Summer Institute of Economics and Finance 2017, Ningbo

International Finance Conference 2016, CUHK–Shenzhen, Tsinghua PBC, CKGSB, Shanghai Jiao

Tong University (SAIF), and HKUST for helpful comments. Liu and Wang acknowledge financial

support from the Hong Kong Research Grants Council (GRF #16504815). The authors have read

the Journal of Finance’s disclosure policy and have no conflicts of interest to disclose.

DOI: 10.1111/jofi.12764

1503

1504 The Journal of Finance R

production (or acquisition) and aggregate economic activities. As the starting

point of our theory, we argue that there exists mutual learning between fi-

nancial markets and the real economy. Their joint information productions

determine both allocative efficiency in the real sector and market efficiency

in the financial sector. As an example, oil producing companies scrutinize oil

futures prices when making their production decisions, while the financial

market studies the financial reports of these producing companies to learn in-

formation when trading on oil futures. This mutual learning creates a strategic

complementarity between information production in the financial sector and

information production in the real sector. A self-fulfilling surge in financial

uncertainty and real uncertainty can naturally arise when both sides produce

little information in anticipation of the other producing little information. At

the same time, aggregate output falls as real allocative efficiency deteriorates.

We formalize the idea in an extended Grossman-Stiglitz (1980) model. Our

key innovation is that we introduce a real sector along the lines of Dixit-Stiglitz

(1977)—in our framework, firms have to make investment decisions under im-

perfect information about two dimensions of uncertainty: their idiosyncratic

productivity (supply) and demand shocks. We start with one firm and one fi-

nancial market in our baseline partial equilibrium model for a given level of

aggregate output. To reduce uncertainty, the firm can learn about its produc-

tivity shock by incurring a cost, but it has to infer its demand shock from infor-

mation provided by the financial market, where speculators (or traders) have

a comparative advantage in acquiring information about the demand shock. In

this context, the financial price is jointly determined by the firm’s information

production and thereby its disclosure about the productivity shock and the in-

formation produced by financial market speculators about the demand shock.

To understand strategic complementarity between these two sources of infor-

mation, first suppose that the firm makes more accurate information disclosure

about its productivity shock. The reduction in uncertainty about the produc-

tivity shock attracts more informed traders and also induces more aggressive

trading of informed traders. Hence, the amount of information provided by the

financial market about the demand shock increases. Conversely, suppose that

for some reason the financial price becomes more informative about the demand

shock. The reduction in uncertainty regarding the demand shock enables the

firm to make better investment decisions and increases its profit multiplica-

tively for every realized productivity shock. This implies that the firm has

stronger incentives to acquire information about its productivity shock.

As the firm’s marginal benefit of acquiring information depends on aggregate

output (in addition to financial price informativeness), the nature of equilib-

rium also depends crucially on the level of aggregate output. At one extreme,

when aggregate output is sufficiently high, the firm’s incentive to acquire in-

formation is already strong enough and acquiring information is a dominant

strategy for the firm. The resulting equilibrium is hence unique in which the

firm produces and discloses more precise information and the financial market

generates a more informative price signal. As a result, both financial uncer-

tainty and real uncertainty are low. At the other extreme, when aggregate

Informational Interdependence 1505

output is sufficiently low, not acquiring information is a dominant strategy for

the firm. Anticipating this, speculators in the financial market also have little

incentive to acquire information about the firm’s demand shock. The equilib-

rium is hence also unique. When aggregate output is in the intermediate range,

however, the economy has two self-fulfilling equilibria. The information pro-

duced by the firm and the information generated by the financial market in one

equilibrium (the “good” equilibrium) are much more precise than those in the

other equilibrium (the “bad” equilibrium). Consequently,a surge in uncertainty

can suddenly strike as a self-fulfilling equilibrium phenomenon.

We next extend the baseline model to a macroeconomic general equilibrium

framework with aggregate production to endogenize the aggregate output. The

level of economic activity and the amount of information available in the econ-

omy are thus jointly determined in general equilibrium. The final consumption

good is produced with a continuum of intermediate capital goods as inputs ac-

cording to a Dixit-Stiglitz production function. Each intermediate capital good

is produced by one firm located on an island in the spirit of Lucas (1972). When

information signals on some islands become noisier, real investment decisions

on those islands become less efficient and consequently aggregate output de-

clines. This causes the aggregate demand faced by other islands to drop. Thus,

incentives to acquire information in the real sector on those other islands are

also reduced, which decreases information acquisition in their financial sectors

as well. Aggregate output therefore declines further, which in turn affects those

islands experiencing the original shock. In short, the complementarity in goods

production due to the Dixit-Stiglitz demand externality across islands gener-

ates further complementarity in information production across islands. As a

result, the complementarity forces for equilibrium multiplicity are strength-

ened in general equilibrium. Similar to the partial equilibrium model, the

economy may feature multiple (two) equilibria. In particular, in general equi-

librium, a self-fulfilling increase in uncertainty is accompanied by a reduction

in investment efficiency and a decline in aggregate output.

We derive four key implications of our macroeconomic model. First, an ad-

verse shock originating in either the real sector or the financial sector that

impairs their ability to conduct information acquisition can have a large im-

pact on the aggregate economy due to the compound feedback loops. Indeed,

both aggregate investment and endogenous aggregate total-factor-productivity

(TFP) are decreasing in information precision. Hence, a small shock in informa-

tion acquisition cost can have a large impact on all three quantities (aggregate

investment, endogenous aggregate TFP, and aggregate output) in the same

direction, particularly when it triggers a self-fulfilling “bad” equilibrium.1Sec-

ond, our model jointly endogenizes the three variables—financial uncertainty,

real uncertainty,and aggregate economic activity—and delivers countercyclical

uncertainty as observed in the data. Third, our model provides an information

contagion channel whereby a shock that directly affects only a small fraction of

islands can generate a global recession on all islands through the endogenous

1Brunnermeier (2009) discusses various other amplification mechanisms of financial markets.