Prozac for depressed states? Effect of mood on local economic recessions

• Published date: 01 April 2020
• Author: Alok Kumar,Vidhi Chhaochharia,George M. Korniotis
• Date: 01 April 2020
• DOI: http://doi.org/10.1002/rfe.1099

Rev Financ Econ. 2020;38:245–274. wileyonlinelibrary.com/journal/rfe

|

245

© 2020 University of New Orleans

1

|

INTRODUCTION

In a recent New York Times article titled “Stuck in Neutral? Reset the Mood”, Robert Shiller conjectures that:

…business recessions are caused by a curious mix of rational and irrational behavior.

Further, he posits that:

Negative feedback cycles, in which pessimism inhibits economic activity, are hard to stop…

Although these are very interesting observations, establishing a causal link between optimism and economic activity is quite

difficult since people's mood and optimism are often influenced by the economic environment. To overcome this difficulty, in this

study, we propose several exogenous mood indicators based on weather, political environment, and sports outcomes, which are

unrelated to the existing economic climate.

Received: 12 November 2019

|

Accepted: 29 December 2019

DOI: 10.1002/rfe.1099

ORIGINAL ARTICLE

Prozac for depressed states? Effect of mood on local economic

recessions

VidhiChhaochharia

|

George M.Korniotis

|

AlokKumar

Department of Finance, Herbert Business

School, University of Miami, Coral Gables,

FL, USA

Correspondence

Alok Kumar, Department of Finance,

Herbert Business School, 514 Jenkins

Building, University of Miami, Coral

Gables, FL 33124, USA.

Email: akumar@miami.edu

Abstract

This paper examines whether people's mood and optimism affect economic activity.

We consider two sets of exogenous proxies for optimism that are unrelated to the

economic environment: (1) weather (average temperature and cloud cover) and (2)

sports and political optimism. We show that economic recessions are weaker and

expansions are stronger in the United States where local individuals are more opti-

mistic. Further, local optimism has a stronger impact on state-level business cycles of

smaller states and regions with low levels of risk sharing. In contrast, the incremental

effects of local optimism are weaker in states where people are younger, more edu-

cated and sophisticated, and socially more connected. States with larger concentra-

tion of minority and urban population also exhibit lower sensitivity to variations in

mood and optimism. Alternative explanations based on the state's industrial compo-

sition, tax environment, migration, seasonal affective disorder (SAD), oil shocks,

and direct economic impact of weather cannot explain these findings.

KEYWORDS

depressed states, economic recession, effect of mood

JEL CLASSIFICATION

E32; R12

246

|

CHHAOCHHARIA et Al.

Further, instead of focusing on the aggregate U.S. economy, we exploit the rich geographical variation in economic con-

ditions and the level of optimism across the United States to investigate the links between them. This choice is motivated

by studies in macroeconomics, which highlight the important differences in the business cycles across the United States. In

one of the earlier studies, MacLaughlin (1930) uses data on 16 U.S. industrial cities for the 1919 to 1921 period and shows

that cities with highly concentrated industries experience more severe cyclical as well as seasonal variations in economic

activity.

More recently, Owyang, Piger, and Wall (2005) examine the timing and characteristics of state-level business cycles and

demonstrate that most states experience distinct recessions and expansions that may be unrelated to the national business cycle.

For example, in the mid-1980s the national economy was expanding but 14 states were experiencing a recession. Further,

Hamilton and Owyang (2011) show that oil-producing and agricultural states tend to experience cycles that can be different

than the economy-wide cycles. They also document substantial differences in the timing of state business cycles where some

states fall and recover from a recession prior to other states.

We recognize these geographical differences in business cycles across the United States and design our empirical tests to

take advantage of these differences. Our hypotheses are inspired by the observations in Shiller (2010). A natural implication

of Robert Shiller's observations is that the mood of individuals within a state would be an important determinant of state-level

business cycles. When local individuals are more optimistic, their spending habits, labor productivity, and entrepreneurial

or other risk-taking activities could be affected, which in turn could have an impact on the local economic environment.

Specifically, we posit that higher levels of local optimism would amplify economic expansions, shorten the length of local

economic recessions, and speed up the recovery process following a recession.

In addition, local optimism is likely to have a stronger impact on state-level business cycles of smaller states and regions with

low levels of risk sharing. Individuals in these states are likely to be more sensitive to income shocks. In contrast, the incremental

effects of local optimism may be weaker in states where people are more educated, sophisticated, and socially more connected.

We test these hypotheses using state-level data on economic activity and optimism. Following Korniotis and Kumar (2013),

we define the state-level economic activity using state-level measures of income growth, relative unemployment, and housing

collateral.1

Further, we consider two sets of exogenous proxies for optimism that are unrelated to the economic environment:

(1) weather (average temperature and cloud cover) and (2) sports and political optimism. These measures allow us to establish a

causal relation between optimism and economic activity as it is hard to argue that changes in the economic environment would

affect these optimism measures.2

Our choice of weather as an indicator of mood and optimism is motivated by the evidence from psychology, which shows

that about 40% of variations in mood can be explained by the local weather (e.g., Persinger and Levesque (1983)). In particular,

high temperatures, more sunshine, and low humidity are associated with positive mood and higher levels of optimism (e.g.,

Cunningham (1979), Howarth and Hoffman (1984)). Recent studies indicate that the relation between weather and mood may

depend upon the time of the year. Specifically, Keller et al. (2005) find that weather affects mood only in the Spring and only

when people spend significant time outdoors. The association between weather and mood also could have a neuro-foundation.

During good weather, the body releases a neuro-transmitter serotonin that makes a person cheerful and more alert. In contrast,

bad weather triggers the release of melatonin that makes people sleepy and depressed.3

To demonstrate that our mood and optimism proxies are appropriate, we examine whether those proxies are correlated with

direct indicators of mood and depression. We find that there are fewer episodes of mental depression and the per-capita con-

sumption of anti-depressants is lower in states in which our mood proxies indicate a positive and more optimistic outlook. This

evidence indicates that our mood and optimism proxies are likely to capture state-level optimism reasonably well.

Previous studies have examined the effect of mood on economic and financial decisions. For example, Saunders (1993)

shows that the local weather in New York City affects the levels of stock market indices through its impact on the mood of

local traders. Using data from different countries, Hirshleifer and Shumway (2003) find that morning sunshine is positively

correlated with daily market returns. Goetzmann and Zhu (2005) shows that the behavior of market-makers may be affected

by New York City weather. In other settings, Kamstra, Kramer, and Levi (2000) show that the Friday to Monday stock market

returns are lower during daylight savings time change weekends. Further, Kamstra, Kramer, and Levi (2003) use country-level

data and show that seasonal variation in daylight influence risk tolerance and aggregate market returns. None of these studies

examine the impact of mood and optimism on the broader economy, which is the main focus of our study.

Our results indicate that economic recessions are weaker and expansions are stronger in the United States where local in-

dividuals are more optimistic. The relation between local mood and the local economy remains strong even when we account

for the direct effects of weather on the local economy. These results are robust and are not period specific. For example, we

find similar (although weaker) results when we use sports and political optimism measures rather than weather-based mood