Dissecting Saving Dynamics
Measuring Credit, Wealth and Precautionary Effects
Christopher Carroll
1
Jiri Slacalek
2
Martin Sommer
3
1
Johns Hopkins University and NBER
[email protected]
2
European Central Bank
[email protected]
3
International Monetary Fund
[email protected]
Presentation at Julis-Rabinowitz Conference
Princeton University, February 2014
US Personal Saving Rate (
s
), 1966–2011
1970
1975
1980
1985
1990
1995
2000
2005
2010
0
2
4
6
8
10
12
14
Literature
I
“Wealth Effects”
I
Modigliani, Klein, MPS model, ...
I
s
t
=
−
0
.
05
m
t
+ other stuff
I
“Precautionary”
I
Carroll (1992)
I
Saving rate rises in recessions
I
∆ log
C
t
+1
strongly related to
E
t
(
u
t
+1
−
u
t
)
I
“Credit Availability”
I
Secular Trend:
I
Parker (2000), Dynan and Kohn (2007), Muellbauer (many
papers)
I
Cyclical Dynamics:
I
Guerrieri and Lorenzoni (2011), Eggertsson and Krugman
(2011), Hall (2011)
Great Recession 2007–2009
I
s
rises by
∼
4 pp
I
Bigger & more persistent increase
than any postwar recession
I
But all three indicators also move a lot:
I
Credit conditions tighten
I
Unemployment Expectations rise
Personal Saving Rate 2007–
↑
−4
−2
0
2
4
Deviation from Start−of−Recession Value in %
0
2
4
6
8
10
12
14
16
18
20
Quarters after Start of Recession
Our Contributions
I
Theory
I
Simple model with transparent role for all 3 channels
I
Qualitative implications of the model
I
“Overshooting”
⇒
possible role for fiscal policy
I
Evidence
I
Quantify the 3 channels
I
Two estimated models of
s
I
Reduced-form—OLS
I
Structural—Nonlinear least squares
I
Conclusions
I
Secular
decline in
s
is almost all from credit
↑
I
Cyclical
movement in
s
is mostly from
w
and
f
Theory `
a la Carroll and Toche (2009)
I
CRRA utility, labor supply
`
, agg wage W, emp status
ξ
:
v
(
m
m
m
t
)
=
max
c
c
c
t
u
(
ccc
t
) +
β
E
t
v
(
m
m
m
t
+1
)
s.t.
m
m
m
t
+1
=
(
m
m
m
t
−
ccc
t
)R +
`
t
+1
W
t
+1
ξ
t
+1
I
ξ
t
+1
∈ {
ξ
u
, ξ
e
}
where
ξ
u
< ξ
e
I
`
and W grow at constant rate
I
Tractability: unemployment shocks are
permanent
I
If
ξ
t
=
ξ
u
then
ξ
t
+1
=
ξ
u
I
Target wealth ˇ
m
exists and is stable:
Target Wealth
m
ˇ
Closed-form solution for target wealth depends on unemployment
risk
f
and generosity of unemployment insurance
ξ
u
:
ˇ
m
=
f
(
f
(+)
, ξ
u
(
−
)
,
preferences
, . . .
)
Consumption After a Wealth Shock
D
m
t
e
+
1
=
0
c
H
m
L
c
t
c
t
+
1
Wealth Shock
Target
c
H
m
L
m
Ç
m
t
m
c
Ç
c
Permanent Rise in
f
Sustainable
c
c
H
m
L
after unemployment rate increase
Target
c
H
m
L
m
Ç
m
Saving Rate After a Permanent Rise in
f
Overshooting
t
Time
s
Ç
t
'
s
t
Credit Easing/Financial Innovation & Deregulation
Orig Target
D
m
t
e
+
1
=
0
Orig
c
H
m
L
New
c
H
m
L
-
h
0.
m
c
ˇ
Net Worth (Ratio to Quarterly Disp Income)
4
4.5
5
5.5
6
6.5
1970
1975
1980
1985
1990
1995
2000
2005
2010
Credit Easing Accumulated (CEA) (`
a la Muellbauer)
Accumulated responses, weighted with debt–income ratio, to:
“Please indicate your
bank’s willingness to make consumer installment loans
now as opposed to three months ago.”
1970 1975 1980 1985 1990 1995 2000 2005 2010 0 0.2 0.4 0.6 0.8 1
f
t
Implied by Michigan U Expectations
I
Regress: ∆
4u
t
+4
=
α
0
+
α
1UExp
t
I
U risk:
f
t
=
u
t
+ ∆
4
u
ˆ
t
+4
I
∆
4u
t
+4
≡
u
t
+4
−
u
t
,
∆
4
ˆ
u
t
+4
≡
fitted values
I
f
t
tracks
but precedes
actual U
UExp
: “How about people out of work during the coming 12 months—do you think
that there will be more unemployment than now, about the same, or less?”
2
4
6
8
10
1970
1975
1980
1985
1990
1995
2000
2005
2010
Reduced-Form Regressions
s
t
=
γ
0
+
γ
m
m
t
+
γ
CEA
CEA
t
+γ
Eu
E
t
u
t
+4
+
γ
t
t
+γ
uC
(
E
t
u
t
+4
×
CEA
t
) +
ε
t
Model
Time
Wealth
CEA
Un Risk
All 3
Baseline
Interact
γ
0
11
.
95
∗∗∗
25
.
20
∗∗∗
9
.
32
∗∗∗
8
.
24
∗∗∗
14
.
90
∗∗∗
15
.
23
∗∗∗
15
.
55
∗∗∗
(0
.
61)
(1
.
73)
(0
.
57)
(0
.
42)
(2
.
56)
(2
.
16)
(2
.
56)
γ
m
−
2
.
61
∗∗∗
−
1
.
12
∗∗∗
−
1
.
18
∗∗∗
−
1
.
37
∗∗∗
(0
.
32)
(0
.
42)
(0
.
35)
(0
.
46)
γ
CEA
−
14
.
14
∗∗∗
−
5
.
47
∗∗∗
−
6
.
12
∗∗∗
−
4
.
60
∗∗∗
(1
.
74)
(1
.
94)
(0
.
57)
(1
.
72)
γ
Eu
0
.
67
∗∗∗
0
.
32
∗∗∗
0
.
29
∗∗∗
0
.
38
∗∗∗
(0
.
05)
(0
.
12)
(0
.
08)
(0
.
11)
γ
t
−
0
.
04
∗∗∗
−
0
.
03
∗∗∗
0
.
04
∗∗∗
−
0
.
05
∗∗∗
−
0
.
00
0
.
00
(0
.
00)
(0
.
00)
(0
.
01)
(0
.
00)
(0
.
01)
(0
.
01)
γ
uC
−
0
.
32
∗∗
(0
.
16)
¯
R
2
0
.
70
0
.
85
0
.
82
0
.
88
0
.
89
0
.
90
0
.
90
F stat p val
0
.
00
0
.
00
0
.
00
0
.
00
0
.
00
0
.
00
0
.
00
DW stat
0
.
30
0
.
69
0
.
50
0
.
86
0
.
94
0
.
93
0
.
98
Fit: Baseline vs Time Trend
2
4
6
8
10
12
1970
1975
1980
1985
1990
1995
2000
2005
2010
s
t
=
γ
1
+
γ
m
m
t
+
γ
CEA
CEA
t
+
γ
Eu
E
t
u
t
+4
+
γ
0
X
t
+
ε
t
Model
Baseline
Uncert
s
t−1Debt
Full Controls
Post-80
IV
γ
m−
1.18
∗∗∗−
1.21
∗∗∗−
0.31
−
0.80
∗∗−
1.30
∗∗∗−
1.50
−
2.02
∗∗∗(0.35)
(0.36)
(0.22)
(0.36)
(0.31)
(1.25)
(0.49)
γ
CEA−
6.12
∗∗∗−
5.97
∗∗∗−
2.87
∗∗∗−
5.40
∗∗∗−
6.24
∗∗∗−
5.00
∗∗−
5.85
∗∗∗(0.57)
(0.65)
(0.53)
(0.73)
(0.63)
(2.00)
(1.17)
γ
Eu0.29
∗∗∗0.28
∗∗∗0.14
∗∗∗0.34
∗∗∗0.12
0.30
∗∗0.08
(0.08)
(0.09)
(0.05)
(0.07)
(0.09)
(0.14)
(0.13)
γ
σ0.26
(0.47)
γ
s0.57
∗∗∗(0.07)
γ
d−
1.91
(1.16)
γ
r0.13
∗∗∗(0.04)
γ
GS−
0.12
(0.08)
γ
CS−
0.31
∗∗(0.14)
γ
0post80−
1.48
(7.90)
γ
mpost800.56
(1.29)
γ
CEApost80−
2.35
(2.13)
Fit: Baseline vs Post-1980
2
4
6
8
10
12
1970
1975
1980
1985
1990
1995
2000
2005
2010
Fit: Baseline vs Full Controls
2
4
6
8
10
12
1970
1975
1980
1985
1990
1995
2000
2005
2010
Structural Estimation—Nonlinear Least Squares
Minimize distance between model-implied
s
t
theor
and actual
s
t
meas
:
ˆ
Θ = arg min
T
X
t
=1
s
t
meas
−
s
t
theor
Θ;
m
t
−
m
ˇ
m
¯
(CEA
t
)
,
f
(
E
t
u
t
+4
)
2
,
where
I
Θ =
{
β,
θ
¯
m
, θ
CEA
,
θ
¯
f
, θ
u
}
I
m
¯
t
= ¯
θ
m
+
θ
CEA
CEA
t
I
f
t
= ¯
θ
f
+
θ
u
E
t
u
t
+4
I
β
: discount factor
Fit of the Structural Model
1970
1975
1980
1985
1990
1995
2000
2005
2010
0
2
4
6
8
10
12
Actual PSR
Fitted PSR
Decomposition of Fitted PSR
Fix
f
t
and CEA
t
at their sample means, back out the implied
s
t
1970
1975
1980
1985
1990
1995
2000
2005
2010
0
2
4
6
8
10
12
Fitted PSR
Fitted PSR excl. Uncertainty
Fitted PSR excl. Uncertainty and CEA
Fit: Structural Model vs Reduced-Form
2
4
6
8
10
12
1970
1975
1980
1985
1990
1995
2000
2005
2010
Actual
Reduced−Form
Structural
PSR Forecasts—In Sample
Great Recession 2007–2010
Variable
Reduced-Form Model
Structural Model
Actual ∆
s
t
γ
m
×
∆
m
t
−
1
.
18
× −
1
.
39 = 1
.
64
−
0
.
97
× −
1
.
39 = 1
.
34
γ
CEA
×
∆CEA
t
−
6
.
12
× −
0
.
11 = 0
.
64
−
6
.
38
× −
0
.
11 = 0
.
67
γ
Eu
×
∆
E
t
u
t
+4
0
.
29
×
4
.
33 = 1
.
24
0
.
32
×
4
.
33 = 1
.
39
PSR Forecasts—Out of Sample
2012–2015
0
2
4
6
8
0
2
4
6
8
2005
2007
2009
2011
2013
2015
Baseline Scenario
Upside Risk Scenario
Downside Risk Scenario
Fitted values of model
(percent of disposable personal income)
Conclusions
I
All three effects present
I
Easier borrowing largely explains secular decline
s
I
Order of importance in Great Recession:
1.
Wealth shock
2.
Labor income risk
3.
Credit tightening
References
Carroll, Christopher D.
(1992): “The Buffer-Stock Theory of Saving: Some Macroeconomic Evidence,”
Brookings Papers on Economic Activity
, 1992(2), 61–156,
http://econ.jhu.edu/people/ccarroll/BufferStockBPEA.pdf
.
Carroll, Christopher D., and Patrick Toche
(2009): “A Tractable Model of Buffer Stock Saving,”
NBER
Working Paper Number 15265
,
http://econ.jhu.edu/people/ccarroll/papers/ctDiscrete
.
Dynan, Karen E., and Donald L. Kohn
(2007): “The Rise in US Household Indebtedness: Causes and
Consequences,” in
The Structure and Resilience of the Financial System
, ed. by Christopher Kent, and Jeremy
Lawson, pp. 84–113. Reserve Bank of Australia.
Eggertsson, Gauti B., and Paul Krugman
(2011): “Debt, Deleveraging, and the Liquidity Trap: A
Fisher-Minsky-Koo Approach,”
Manuscript, NBER Summer Institute
.
Guerrieri, Veronica, and Guido Lorenzoni
(2011): “Credit Crises, Precautionary Savings and the Liquidity
Trap,”
Manuscript, MIT Department of Economics
.
Hall, Robert E.
(2011): “The Long Slump,” AEA Presidential Address, ASSA Meetings, Denver.
Parker, Jonathan A.
(2000): “Spendthrift in America? On Two Decades of Decline in the U.S. Saving Rate,”
in
NBER Macroeconomics Annual 1999
, ed. by Ben S. Bernanke, and Julio J. Rotemberg, vol. 14, pp.
317–387. NBER.
Alternative Measures of Credit Availability
.6
.7
.8
.9
1
Abiad et al. Index of Financial Liberalization
0
.5
1
1.5
CEA/Debt−Income Ratio
1970
1975
1980
1985
1990
1995
2000
2005
2010
Assumptions/Scenarios for Out-of-Sample Forecasts
Sources: Haver Analytics and authors' estimates. 400 450 500 550 600 650 700 400 450 500 550 600 650 700 2005 2007 2009 2011 2013 2015 Baseline scenario Upside risk scenario Downside risk scenario (percent of disposable personal income)
4 6 8 10 12 4 6 8 10 12 2005 2007 2009 2011 2013 2015 Baseline scenario Upside risk scenario Downside risk scenario Unemployment expectations (percent of labor force)
0.7 0.8 0.9 1.0 1.1 1.2 1.3 0.7 0.8 0.9 1.0 1.1 1.2 1.3 2005 2007 2009 2011 2013 2015 Baseline scenario Upside risk scenario Downside risk scenario
0 2 4 6 8 0 2 4 6 8 2005 2007 2009 2011 2013 2015 Baseline Scenario Upside Risk Scenario Downside Risk Scenario Fitted values of model (percent of disposable personal income)
Household net wealth
Unemployment rate
Assumptions/Scenarios for Out-of-Sample Forecasts
Sources: Haver Analytics and authors' estimates. 400 450 500 550 600 650 700 400 450 500 550 600 650 700 2005 2007 2009 2011 2013 2015 Baseline scenario Upside risk scenario Downside risk scenario (percent of disposable personal income)
4 6 8 10 12 4 6 8 10 12 2005 2007 2009 2011 2013 2015 Baseline scenario Upside risk scenario Downside risk scenario Unemployment expectations (percent of labor force)
0.7 0.8 0.9 1.0 1.1 1.2 1.3 0.7 0.8 0.9 1.0 1.1 1.2 1.3 2005 2007 2009 2011 2013 2015 Baseline scenario Upside risk scenario Downside risk scenario
0 2 4 6 8 0 2 4 6 8 2005 2007 2009 2011 2013 2015 Baseline Scenario Upside Risk Scenario Downside Risk Scenario Fitted values of model (percent of disposable personal income)
