Understanding the Deep Contraction and Slow
Recovery of the U.S. Economy
Robert E. Hall
Hoover Institution and Department of Economics
Stanford University
24 September 2015
IFS and Bank of England Conference
Understanding the Great Recession: From Micro to Macro
·
1
The plan
Quick review of the slump following the financial crisis of 2008
2
The plan
Quick review of the slump following the financial crisis of 2008
Ideas about driving forces and propagation mechanisms
2
The plan
Quick review of the slump following the financial crisis of 2008
Ideas about driving forces and propagation mechanisms
A model constructed to study propagation
·
2
The recent slump was similar to earlier ones
Peak year
Peak rate
Ratio of later unemployment rate to peak
rate, by number of years later
1
2
3
4
1975
8.5
0.91
0.84
0.72
0.69
1982
9.7
0.99
0.77
0.74
0.72
1992
7.5
0.92
0.81
0.75
0.72
2010
9.6
0.93
0.84
0.77
0.65
3
Real GDP, 2000-2014, Billions of 2009 Dollars
18,000
16,000
14,000
12,000
10,000
8,000
6,000
2000
2002
2004
2006
2008
2010
2012
2014
4
Real Consumption Expenditure, 2000-2014, Billions
of 2009 Dollars
12,000
11,000
10,000
9,000
8,000
7,000
6,000
2000
2002
2004
2006
2008
2010
2012
2014
5
Employment, 2000-2014, Thousands of Workers
150,000
140,000
130,000
120,000
110,000
100,000
90,000
2000
2002
2004
2006
2008
2010
2012
2014
6
Unemployment, 2000-2014, Percent of Labor Force
12
10
8
6
4
2
0
2000
2002
2004
2006
2008
2010
2012
2014
7
Percent of Working-Age Population in the Labor
Force, 2000-2014
68
67
66
65
64
63
62
61
60
2000
2002
2004
2006
2008
2010
2012
2014
8
Average Real Earnings per Household, 2009
Dollars, 1990-2014
80,000
70,000
60,000
50,000
40,000
30,000
20,000
10,000
0
1990
1993
1996
1999
2002
2005
2008
2011
9
Index of Capital Services, 2007 = 1, 2000-2014
1.2
1.1
1.0
0.9
0.8
0.7
0.6
0.5
0.4
2000
2002
2004
2006
2008
2010
2012
2014
10
Index of Total Factor Productivity, 2007 = 1,
2000-2014
1.10
1.05
1.00
0.95
0.90
0.85
0.80
0.75
0.70
2000
2002
2004
2006
2008
2010
2012
2014
11
Labor share
1.30
1.25
1.20
1.15
1.10
1.05
1.00
0.95
0.90
0.85
1989
1992
1995
1998
2001
2004
2007
2010
2013
12
Business Earnings as a Ratio to the Value of Capital
0.25
0.20
0.15
0.10
0.05
0.00
2000
2002
2004
2006
2008
2010
2012
13
Driving forces
Some potentially exogenous
14
Driving forces
Some potentially exogenous
Others clearly induced by a process outside the model, such as rising discounts and
frictions resulting from collapse of house prices
·
14
Driving force #1: Labor-force participation
Not generally considered in the fluctuations literature
15
Driving force #1: Labor-force participation
Not generally considered in the fluctuations literature
Decline of about 3 percentage points after 2007; not the result of the slack labor
market
15
Driving force #1: Labor-force participation
Not generally considered in the fluctuations literature
Decline of about 3 percentage points after 2007; not the result of the slack labor
market
Continuation of a trend starting in 2000, not the result of demographic shifts
·
15
Driving force #2: Capital wedge
Difference between the return to capital and the safe real interest rate
16
Driving force #2: Capital wedge
Difference between the return to capital and the safe real interest rate
Comprises agency frictions, risk premium, and taxes
·
16
Measuring the wedge
qt = κ
kt
−1 +1
kt−1
17
Measuring the wedge
qt = κ
1 + rk,t =
kt
−1 +1
kt−1
1
πt
+ (1 − δt )qt+1 pk,t+1
qt pk,t kt
17
Measuring the wedge
qt = κ
1 + rk,t =
kt
−1 +1
kt−1
1
πt
+ (1 − δt )qt+1 pk,t+1
qt pk,t kt
ft = rk,t − rf,t
·
17
The capital wedge for two values of the adjustment
cost κ
20
20
18
18
16
16
14
14
12
12
10
10
8
8
6
6
4
4
2
2
0
0
1990
1993
1996
1999
2002
2005
2008
2011
1990
(a) κ = 0
1993
1996
1999
2002
2005
2008
2011
(b) κ = 2
·
18
Driving force #3: Extra discount
Discounts for risky payoffs seem to rise far more than the return to capital during
and after crises
·
19
The S&P Risk Premium, 1960 through 2012
30
25
20
15
10
5
0
‐5
1960
1965
1970
1975
1980
1985
1990
1995
2000
2005
2010
20
BBB-Treasury Bond Spread
5.0
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
2000
2002
2004
2006
2008
2010
2012
2014
21
Driving force #4: Total factor productivity
Fernald concludes that the decline was not the result of the crisis
·
22
Driving force #5: Shift in product demand induced
by contraction in lending
Shifts from different sources all result in “multiplier” responses
23
Driving force #5: Shift in product demand induced
by contraction in lending
Shifts from different sources all result in “multiplier” responses
No important shift in government purchases, the usual variable
23
Driving force #5: Shift in product demand induced
by contraction in lending
Shifts from different sources all result in “multiplier” responses
No important shift in government purchases, the usual variable
But a major decline in household consumption as a result of tightening lending
standards and declining collateral
23
Driving force #5: Shift in product demand induced
by contraction in lending
Shifts from different sources all result in “multiplier” responses
No important shift in government purchases, the usual variable
But a major decline in household consumption as a result of tightening lending
standards and declining collateral
Some part of the decline in business investment also resulting from tightening
standards
·
23
Driving force #6: Product-market wedge
In the New Keynesian view, the wedge rises in slumps
24
Driving force #6: Product-market wedge
In the New Keynesian view, the wedge rises in slumps
Another view: Financially constrained firms raise markups
·
24
Propagation mechanisms
Propagation occurs when the effect of a driving force lasts longer than the force
itself
·
25
Mechanism #1: Depletion of the capital stock
This turns out to be the leading source of propagation
·
26
Mechanism #2: Unemployment dynamics
Because job-finding rates remained high even during the recent slump, leading
labor-market models seem to imply that little propagation occurs because of the
time it takes displaced workers to find new jobs
27
Mechanism #2: Unemployment dynamics
Because job-finding rates remained high even during the recent slump, leading
labor-market models seem to imply that little propagation occurs because of the
time it takes displaced workers to find new jobs
But studies of the experiences of displaced workers shows length periods of lower
employment
27
Mechanism #2: Unemployment dynamics
Because job-finding rates remained high even during the recent slump, leading
labor-market models seem to imply that little propagation occurs because of the
time it takes displaced workers to find new jobs
But studies of the experiences of displaced workers shows length periods of lower
employment
A high incidence of very brief interim jobs seems to be at least part of the
reconciliation
27
Mechanism #2: Unemployment dynamics
Because job-finding rates remained high even during the recent slump, leading
labor-market models seem to imply that little propagation occurs because of the
time it takes displaced workers to find new jobs
But studies of the experiences of displaced workers shows length periods of lower
employment
A high incidence of very brief interim jobs seems to be at least part of the
reconciliation
The slow but steady decline of high rates of long-term unemployment is consistent
with this view
·
27
Mechanism #3: The Zero Lower Bound
At the lower bound, the real interest rate is minus the expected rate of inflation
28
Mechanism #3: The Zero Lower Bound
At the lower bound, the real interest rate is minus the expected rate of inflation
In the Great Depression, expected inflation turned negative, so real rates reached
high levels
28
Mechanism #3: The Zero Lower Bound
At the lower bound, the real interest rate is minus the expected rate of inflation
In the Great Depression, expected inflation turned negative, so real rates reached
high levels
In the recent slump, expected inflation remained roughly constant
28
Mechanism #3: The Zero Lower Bound
At the lower bound, the real interest rate is minus the expected rate of inflation
In the Great Depression, expected inflation turned negative, so real rates reached
high levels
In the recent slump, expected inflation remained roughly constant
Even so, the bound elevated unemployment for as long as it lasted, more than six
years
·
28
Inflation Expectations and Forecasts
4.5
4.0
3.5
Michigan
3.0
2.5
Professional forecasters
TIPSs
2.0
1.5
1.0
0.5
0.0
2003
2005
2007
2009
2011
2013
29
Effects of the ZLB
Almost all models portray the ZLB as opening a gap between output supply and
demand and labor supply and demand
30
Effects of the ZLB
Almost all models portray the ZLB as opening a gap between output supply and
demand and labor supply and demand
Gapology seems realistic but leaves mysteries about how markets respond to
congestion resulting from excess supply
30
Effects of the ZLB
Almost all models portray the ZLB as opening a gap between output supply and
demand and labor supply and demand
Gapology seems realistic but leaves mysteries about how markets respond to
congestion resulting from excess supply
And it’s hard to reconcile gap-based unemployment with DMP-based
unemployment
30
Effects of the ZLB
Almost all models portray the ZLB as opening a gap between output supply and
demand and labor supply and demand
Gapology seems realistic but leaves mysteries about how markets respond to
congestion resulting from excess supply
And it’s hard to reconcile gap-based unemployment with DMP-based
unemployment
The ZLB propagated the declines in product demand resulting from financial
events—the multiplier for these declines was much higher than normal, according
to the demand-gap model
·
30
Model suited to studying persistence
Medium term; 2-year periods
31
Model suited to studying persistence
Medium term; 2-year periods
Exact stochastic solution
31
Model suited to studying persistence
Medium term; 2-year periods
Exact stochastic solution
Non-stationary; models unfiltered data
·
31
Properties
Four-state Markov process for the 6 driving forces
32
Properties
Four-state Markov process for the 6 driving forces
TFP and labor force evolve as trended random walks with increments a function of
the 4-valued random state
32
Properties
Four-state Markov process for the 6 driving forces
TFP and labor force evolve as trended random walks with increments a function of
the 4-valued random state
Levels of other driving forces functions of the state
32
Properties
Four-state Markov process for the 6 driving forces
TFP and labor force evolve as trended random walks with increments a function of
the 4-valued random state
Levels of other driving forces functions of the state
States constructed from principal components of the driving forces
·
32
The States of the Model
State
TFP
growth
Discount
Laborforce
growth
Wedge
Periods in state
1
H
L
L
L
1952-53, 1954-55, 1958-59, 1962-63, 198889, 1990-91, 1996-97, 1998-99
2
H
L
H
L
1956-57, 1960-61, 1964-65, 1966-67, 196869, 1970-71, 1986-87, 2000-1
3
L
H
L
H
1982-83, 1992-93, 1994-95, 2004-5, 20067, 2008-9, 2010-11, 2012-13
4
L
H
H
H
1972-73, 1974-75, 1976-77, 1978-79, 198081, 1984-85, 2002-3
33
Transition Matrix and Ergodic Distribution
To state
From
state
Ergodic
probability
1
2
3
4
1
0.38
0.50
0.13
0.00
0.21
2
0.38
0.38
0.00
0.25
0.22
3
0.14
0.00
0.71
0.14
0.33
4
0.00
0.14
0.29
0.57
0.24
34
Persistence of the States
0.9
Excess probability of being in state
0.8
0.7
1
2
4
3
0.6
0.5
0.4
0.3
0.2
0.1
0
0
2
4
6
8
10
Years
35
Responses of Driving Forces to Their Own Shocks
0.014
0.045
0.040
0.010
0.005
0.012
0.035
0.000
0.020
0.015
Expected value
Expected value
0.025
0.008
0.006
‐0.005
‐0.010
‐0.015
‐0.020
0.004
0.010
‐0.025
0.002
0.005
0.000
‐0.030
0.000
0
2
4
6
‐0.035
0
8
2
4
Years
6
8
Years
(a) Productivity
0
2
4
(c) Discount
0.005
0.000
‐0.005
‐0.010
‐0.015
‐0.020
0
2
4
6
8
Years
(d) Wedge
6
Years
(b) Labor force
Expected value
Expected value
0.010
0.030
36
8
Model specification
One-sector growth model with adjustment costs for investment
37
Model specification
One-sector growth model with adjustment costs for investment
Time to build requires that investment and hiring are determined in the period
before they take effect
37
Model specification
One-sector growth model with adjustment costs for investment
Time to build requires that investment and hiring are determined in the period
before they take effect
In the asset-pricing equations, marginal utility one period ahead is elevated by the
discount driving force, so investment and hiring decisions are made with elevated
discount
37
Model specification
One-sector growth model with adjustment costs for investment
Time to build requires that investment and hiring are determined in the period
before they take effect
In the asset-pricing equations, marginal utility one period ahead is elevated by the
discount driving force, so investment and hiring decisions are made with elevated
discount
The zero-profit condition for recruiting effort in the labor market indexes the
future benefit to the employer by the marginal revenue product of labor
·
37
The zero lower bound in the model
When the ZLB binds, the DMP zero-profit condition drops out
38
The zero lower bound in the model
When the ZLB binds, the DMP zero-profit condition drops out
Instead, the asset pricing condition for the safe real rate holds that rate at a
constant
·
38
Productivity Shock
9
9
8
8
Consumption
Consumption
7
Response to initial shock
Response to initial shock
7
6
Capital
5
4
3
2
Employment
6
4
3
2
1
1
0
0
‐1
Capital
5
Employment
‐1
0
2
4
6
8
0
Years
2
4
6
8
Years
(a) Market rate
(b) Fixed rate
·
39
9
9
7
7
Response to initial shock
Response to initial shock
Discount Shock
5
3
1
Employment
‐1
Capital
‐3
Consumption
0
2
4
6
Employment
5
3
Consumption
1
Capital
‐1
‐3
8
0
Years
2
4
6
8
Years
(a) Market rate
(b) Fixed rate
·
40
4.5
4.5
4.0
4.0
3.5
3.5
Consumption
Consumption
Response to initial shock
Response to initial shock
Labor-Force Shock
3.0
2.5
Capital
2.0
Employment
1.5
1.0
Capital
3.0
2.5
2.0
Employment
1.5
1.0
0.5
0.5
0.0
0.0
‐0.5
‐0.5
0
2
4
6
0
8
2
4
6
8
Years
Years
(a) Market rate
(b) Fixed rate
·
41
Financial-Friction Shock
0.5
0.5
Employment
0.0
Employment
Capital
Response to initial shock
Response to initial shock
0.0
‐0.5
Consumption
‐1.0
‐1.5
‐2.0
‐0.5
‐1.0
Capital
‐1.5
‐2.0
‐2.5
‐2.5
‐3.0
‐3.0
0
2
4
6
8
Consumption
0
Years
2
4
6
8
Years
(a) Market rate
(b) Fixed rate
·
42
Product-Market Wedge Shock
0.0
Capital
‐0.2
Consumption
‐0.3
Employment
‐0.1
Response to initial shock
Response to initial shock
0.0
Employment
‐0.1
‐0.4
‐0.5
Capital
‐0.2
‐0.3
Consumption
‐0.4
‐0.5
‐0.6
‐0.6
0
2
4
6
8
0
Years
2
4
6
8
Years
(a) Market rate
(b) Fixed rate
·
43
Product-Market Demand Shock
1.3
Response to initial shock
Response to initial shock
1.3
0.8
0.3
Employment
‐0.2
0.8
Employment
Consumption
0.3
Capital
‐0.2
Capital
Consumption
‐0.7
‐0.7
0
2
4
6
8
0
Years
2
4
6
8
Years
(a) Market rate
(b) Fixed rate
·
44
Contributions of driving forces to standard
deviations of 2-year log changes in the endogenous
variables
Driving force
Interest
rate
Endogenous variable
Consumption
Capital
Employment
Market
0.015
0.019
0.009
Fixed
0.016
0.021
0.030
TFP
Market
0.014
0.019
0.009
Fixed
0.031
0.005
0.096
Discount
Market
0.007
0.008
0.017
Fixed
0.010
0.010
0.026
Labor force
Market
0.002
0.002
0.001
Fixed
0.009
0.007
0.034
Financial friction
Market
0.001
0.001
0.002
Fixed
0.002
0.001
0.006
Product market wedge
Market
0.001
0.002
0.001
Fixed
0.004
0.001
0.014
0.030
0.023
0.032
Government purchases
Actual data
45
Propagation Ratios by Driving Force, Interest Rate
Regime, and Endogenous Variable
Driving force
Interest
rate
Endogenous variable
Consumption
Capital
Employment
Market
3.5
449.2
2.4
Fixed
1.5
2.5
0.1
TFP
Market
1.8
0.9
0.3
Fixed
0.3
0.6
0.1
Market
3.3
17.5
1.8
Fixed
1.5
2.5
0.8
Market
2.3
0.8
0.3
Discount
Labor force
Financial friction
Fixed
0.8
0.8
0.2
Market
1.1
0.9
0.3
Fixed
0.8
0.8
0.1
Product market wedge
Market
1.9
0.9
0.3
Fixed
0.3
0.7
0.1
Government purchases
46