Transcript Animal Spirits and Economic Fluctuations

ANIMAL SPIRITS
AND ECONOMIC FLUCTUATIONS
SHI FANG
Adviser: Prof. Peter Matthews
ECON 700 Senior Research
Introduction/Motivation

Animal Spirits and Economic Activity
John Maynard Keynes
 Irrational human emotion or sentiments
that are not the outcome “of a weighted
average of quantitative benefits
multiplied by quantitative probabilities.”
 Irrational Confidence vs. Rational
Confidence
 Past recessions in the U.S.

Introduction/Motivation
Early 2000s Recession
120
Confidence Index
100
80
60
Business
40
Consumer
20
0
Quaters
Introduction/Motivation
1990s Recession
100
Confidence Index
90
80
70
60
50
40
Business
30
Consumer
20
10
0
Quaters
Introduction/Motivation
Current Recession
100
90
Confidence Index
80
70
60
50
40
Business
30
Consumer
20
10
0
06Q4 07Q1 Q2
Q3
Q4 08Q1 Q2
Quarters
Q3
Q4 09Q1
Literature Review

Matsusaka and Sbordone (1995)



Chauvet and Guo (2003)



“Consumer Confidence and Economic Fluctuations.”
Self-fulfilling pessimism: an important independent factor in
affecting aggregate output.
“Sunspots, Animal Spirits, and Economic Fluctuations.”
“Animal spirits” have played a nontrivial role in the 1969-1970,
the 1973-1975, and the 1981-1982 recessions.
Akerlof and Shiller (2009)


Animal Spirits: How Human Psychology Drives the Economy, and
Why It Matters for Global Capitalism.
Homo economicus is an unrealistic notion. Emotions due to
noneconomic motivations should be taken into account.
Data

Consumer Confidence
University of Michigan Consumer Sentiment Index
 The Conference Board Consumer Confidence Index


Business Confidence


The Conference Board CEO Confidence Survey
Variables capturing economic fundamentals
Bureau of Economic Analysis (GDP, Personal Income, etc)
 The Federal Reserve (Selected Interest Rates)
 Moody’s Dismal Scientist

Data Summary

Time Series Sample Period
1976 Q2 --- 2009 Q1
 132 Quarters


Selected Summary Statistics
Variable
Mean
Std. Dev.
Min
Max
Consumer(UM)
87.00
12.30
54.4
110.1
Consumer(CB)
96.17
22.75
26.9
142.5
Business
53.51
10.45
24
76
Real GDP
8871.42
2566.83
5128.9
13415.3
3-month TB
5.75
3.12
0.21
15.05
VAR Model
where
𝑦𝑡 = 𝑦1𝑡 , 𝑦2𝑡 , … , 𝑦𝐾𝑡 ′ is a 𝐾 × 1 random vector,
𝐴1 , … , 𝐴𝑝 are 𝐾 × 𝐾 matrices of parameters,
𝑣 is a 𝐾 × 1 vector of intercept parameters, and
𝑢𝑡 represents the innovations and is assumed to be
white noise, such that
𝐸 𝑢𝑡 = 0,
𝐸 𝑢𝑡 𝑢𝑡′ = cross-equation error variance-covariance
matrix , and
𝐸 𝑢𝑡 𝑢𝑠′ = 0 for 𝑡 ≠ 𝑠
VAR Model

Vector Autoregression


A n-equation, n-variable linear model in which each
variable is in turn explained by its own lagged values, plus
current and past values of the remaining n-1 variables.
Model Selection





Akaike's information criterion (AIC)
Schwarz's Bayesian information criterion (SBIC)
Hannan and Quinn information criterion (HQIC)
Autocorrelation
Stability condition
VAR Model

Best Specification (4-Variable with 4 Lags)


Consumer confidence (UM), business confidence, 3month Treasury Bill interest rate, and first difference in
log real GDP
Granger Causality Tests
Dependent Variable in Regression
Regressor
Consumer
Business
GDP
Interest Rate
Consumer
0.00
0.52
0.25
0.12
Business
0.13
0.00
0.16
0.17
GDP
0.74
0.02**
0.00
0.02**
Interest Rate
0.12
0.01**
0.06*
0.00
All
0.16
0.00**
0.00**
0.03**
Results

Impulse Response Functions (IRF)
IRF trace out the response of current and future values
of each of the variables to a one-unit increase in the
current value of one of the VAR errors/innovations,
assuming that this error returns to zero in subsequent
periods and that all other errors are equal to zero.
 Structurally interpretable IRF obtained by
orthogonalized innovations via Cholesky decomposition
 Order: GDP, Interest Rate, Business Confidence,
Consumer Confidence

Impulse Response

Suppose that the VAR is stable, we can derive the vector movingaverage representation of the VAR.
∞
𝒚𝒕 = 𝝁 +
𝚽𝐢 𝐮𝐭−𝐢
𝒊=𝟎
where
𝜇 is the 𝐾 × 1 time-invariant mean of the process, and
Φi are 𝐾 × 𝐾 matrices of parameters.
The process by which the variables in 𝑦𝑡 fluctuate about their time-invariant
means, 𝜇, is completely determined by the parameters in Φi and the (infinite)
past history of the independent and identically distributed shocks or
innovations, 𝑢𝑡 , 𝑢𝑡−1 , …
Impulse Response
Orthogonal IRF
Impulse (10-unit Business), Response (GDP)
% Growth in Real GDP
3
2.5
2
1.5
OIRF
1
Lower
0.5
Upper
0
-0.5
-1
0
1
2
3
4
5
6
7
Quarter
8
9
10 11 12
Impulse Response
Cumulative Orthogonal IRF
Impulse (10-unit Business), Response (GDP)
% Growth in Real GDP
20
15
10
COIRF
5
Lower
Upper
0
0
1
2
3
4
5
6
7
Quarter
8
9
10 11 12
Impulse Response
Orthogonal IRF
Impulse (10-unit Consumer), Response (GDP)
% Growth in Real GDP
2.5
2
1.5
OIRF
1
0.5
Lower
0
Upper
-0.5
-1
0
1
2
3
4
5
6
7
Quarter
8
9
10 11 12
Impulse Response
Cumulative Orthogonal IRF
Impulse (10-unit Consumer), Response (GDP)
% Growth in Real GDP
14
12
10
8
COIRF
6
4
Lower
2
Upper
0
-2
0
1
2
3
4
5
6
7
Quarter
8
9
10 11 12
Conclusion



Animal spirits in business expectations has real,
significant macroeconomic consequences.
Animal spirits in consumer sentiment, however, has a
relatively less significant impact in affecting
macroeconomic activities.
Limitations of the model due to Cholesky
decomposition.
Questions/Discussions