Inflation Dynamics During the Financial Crisis
S. Gilchrist1
R. Schoenle2
1 Boston
J. Sim3
E. Zakrajšek3
University and NBER
2 Brandeis
3 Federal
University
Reserve Board
Second BIS Research Network Meeting
“Macroeconomics and Financial Markets”
Basel, Switzerland
March 10–11, 2015
D ISCLAIMER: The views expressed are solely the responsibility of the authors and should not be interpreted as reflecting the views of the
Board of Governors of the Federal Reserve System or of anyone else associated with the Federal Reserve System.
M OTIVATION
In spite of substantial and persistent economic slack, U.S. saw only mild
disinflation during the “Great Recession” and subsequent slow recovery.
Hall [2011]; Ball & Mazumder [2011]; King & Watson [2012]
What accounts for this “missing deflation?”
◮
Unanchored expectations?
Coibon & Gorodnichenko [2015]
◮
Unusual labor market developments (short- vs. long-term unemployed)?
Gordon [2013]; Krueger et al. [2014]
◮
Actually, there is no missing deflation puzzle.
Del Negro et al. [2015]; Christiano et al. [2015]
O UR PAPER
Can interaction of customer markets and financial frictions help explain
inflation dynamics during the 2007–09 crisis?
Empirics:
◮
◮
Merge good-level prices in the Producers Price Index (PPI) to producers’
income and balance sheet data from Compustat.
Analyze how differences in firms’ internal liquidity positions affected
their price-setting behavior during the crisis.
Theory:
◮
◮
Develop a dynamic GE model that embeds financial frictions in a
customer-markets framework.
Analyze inflation and output dynamics in response to demand and
financial shocks.
DATA S OURCES AND M ETHODS
Monthly good-level price data underlying the PPI.
Nakamura & Steinsson [2008]; Goldberg & Hellerstein [2009]; Bhattarai & Schoenle [2010]
Match about 600 PPI respondents to their income and balance sheet data
from Compustat.
◮
◮
Sample period: Jan2005–Dec2012
Matched PPI-Compustat sample is representative of broader
macroeconomic trends.
Inflation by financial and product market characteristics:
◮
Liquidity ratio (LIQ) ⇒ financial frictions
Campello et al. [2011]
◮
SG&A expense ratio (SGAX) ⇒ customer markets
Goriou & Rudanko [2011]
I NDUSTRY-A DJUSTED P RODUCER P RICE I NFLATION
By financial characteristics
Percentage points
4
3-month moving average
2
0
-2
Low liquidity firms
High liquidity firms
-4
-6
2005
2006
2007
2008
2009
2010
2011
2012
N OTE : Weighted-average inflation relative to industry (2-digit NAICS) inflation (seasonally adjusted monthly rate).
I NDUSTRY-A DJUSTED P RODUCER P RICE I NFLATION
By product market characteristics
Percentage points
3
3-month moving average
Low SGAX firms
High SGAX firms
2
1
0
-1
-2
-3
-4
2005
2006
2007
2008
2009
2010
2011
2012
N OTE : Weighted-average inflation relative to industry (2-digit NAICS) inflation (seasonally adjusted monthly rate).
I NDUSTRY-A DJUSTED P RODUCER P RICES
By financial and product market characteristics as of 2006
Index (Jan. 2008 = 100)
Index (Jan. 2008 = 100)
110
Monthly
Low liquidity firms
High liquidity firms
110
Monthly
105
105
100
100
95
95
90
85
Low SGAX firms
High SGAX firms
80
2008
2009
2010
2011
2012
(a) By liquidity ratio
N OTE : Cumulative weighted-average industry-adjusted inflation rates.
90
85
2008
2009
2010
2011
(b) By SGAX ratio
2012
I NDUSTRY-A DJUSTED P RODUCER P RICE I NFLATION
By financial characteristics and durability of output
Percentage points
4
3-month moving average
2
0
-2
Low liquidity firms - durable goods
High liquidity firms - durable goods
Low liquidity firms - nondurable goods
High liquidity firms - nondurable goods
-4
-6
-8
-10
2005
2006
2007
2008
2009
2010
2011
2012
N OTE : Weighted-average inflation relative to industry (2-digit NAICS) inflation (seasonally adjusted monthly rate).
P RICE -S ETTING B EHAVIOR D URING THE C RISIS
Multinomial logit (3-month-ahead directional price change):

 +
0 = Λ(LIQj,t , SGAXj,t , Xj,t ; β1 , β2 , θ)
Pr sgn(∆3 pi,j,t+3 ) =

−
Inflation regression (3-month-ahead):
3m
πi,j,t+3
= β1 LIQj,t + β2 SGAXj,t + θ ′ Xj,t + ηj + ui,j,t+3 ,
Estimation:
◮
◮
Coefficients β1 and β2 are allowed to switch in 2008.
4-quarter rolling window
D IRECTIONAL P RICE C HANGE M ARGINAL E FFECTS
With respect to liquidity ratio (4-quarter rolling window estimates)
Probability of a price increase
Probability of a price decrease
0.4
0.4
Estimate
+/- 2 SEs
0.2
0.3
0.0
0.2
-0.2
0.1
-0.4
0.0
-0.6
-0.1
Estimate
+/- 2 SEs
-0.8
2006
2008
2010
2012
(c) Marginal effect of liquidity ratio
-0.2
2006
2008
2010
2012
(d) Marginal effect of liquidity ratio
Implications: 2 std. deviation difference in LIQ ⇒ 11 pps. difference in
probability of a price increase.
I NFLATION E FFECTS
With respect to liquidity ratio (4-quarter rolling window estimates)
0.04
0.02
0.00
-0.02
-0.04
Estimate
+/- 2 SEs
-0.06
-0.08
2006
2007
2008
2009
2010
2011
2012
Implications: 2 std. deviation difference in LIQ ⇒ 4 pps. difference in
annualized inflation.
GE M ODEL
Customer markets imply that firms trade off current profits for future
market share.
Phelps & Winter [1970]; Bils [1989]
Financial market frictions imply that firms discount the future more
when demand is low—and therefore maintain high markups.
Gottfries [1991]; Chevalier & Scharfstein [1996]
Embed this intuition into a GE model with nominal price rigidities.
P REFERENCES : “D EEP H ABITS ”
Ravn, Schmitt-Grohe and Uribe [2006]
Problem of household j ∈ [0, 1]:
max Et
∞
X
j
β s U(xt+s
− ψt+s , hjt+s )
s=0
Habit-adjusted consumption bundle:

xtj ≡ 
◮
Z
1
0
cjit
sθi,t−1
!1− 1
η

di
1
1
1− η
;
Law of motion for the external habit:
sit = ρsi,t−1 + (1 − ρ)cit ;
◮
θ < 0 and η > 0
ψt = demand shock
0<ρ<1
T ECHNOLOGY
Continuum of monopolistically competitive firms producing a variety of
differentiated goods indexed by i ∈ [0, 1].
Production function (labor input only):
α
At
yit =
− φi ;
hit
ait
◮
◮
◮
0<α≤1
At = aggregate technology level
ait = i.i.d. idiosyncratic technology shock with
log ait ∼ N(−0.5σ 2 , σ 2 )
φi = fixed operating costs
Baseline case: homogeneous firms (φi = φ, ∀i)
F RICTIONS AND M ONETARY P OLICY
Frictions:
◮
Nominal rigidities:
Rotemberg [1982]
γp
2
◮
Pit
− π̄
Pi,t−1
2
ct =
γp
2
πt
pit
pi,t−1
− π̄
2
ct ;
pit ≡
Costly external equity financing:
Myers & Majluf [1984]; Gomes [2001]; Stein [2003]
• dilution cost (0 < ϕt < 1) ⇒ 1$ of issuance brings in (1 − ϕt )$
Monetary authorities:
rt = (1 + rt−1 )
◮
τr
(1 + r̄)
π τπ y τy 1−τr
t
t
π∗
y∗t
− 1.
Baseline case: central bank cares only about inflation (τy = 0)
Pit
Pt
T IMING AND E QUILIBRIUM
Within-period sequence of events:
(1) Aggregate information arrives in the morning
(2) Post prices based on aggregate information
(3) Take orders, plan production based on expected marginal cost
(4) Idiosyncratic shock ait realized after orders have been taken
(5) Meet demand based on originally posted prices and orders
Risk-neutrality, timing, and i.i.d. shocks imply symmetric equilibrium:
◮
◮
All firms choose identical price (pit = pt = 1) and scale (cit = ct )
Symmetry does not apply to hit and dit .
L OG -L INEARIZED P HILLIPS C URVE
Standard New Keynesian model
π̂t
#
"
∞
X
ω(η − 1)
= −
µ̂t + Et
χδ̃ s−t+1 µ̂s+1 + βEt [π̂t+1 ]
γp
s=t
#
"
∞
X
1
s−t+1
χδ̃
(ξˆt − ξˆs+1 ) − β̂t,s+1
+ [η − ω(η − 1)] Et
γp
s=t
L OG -L INEARIZED P HILLIPS C URVE
The role of “deep habits”
π̂t
#
"
∞
X
ω(η − 1)
= −
χδ̃ s−t+1 µ̂s+1 + βEt [π̂t+1 ]
µ̂t + Et
γp
s=t
"
#
∞
X
1
s−t+1
+ [η − ω(η − 1)] Et
χδ̃
(ξˆt − ξˆs+1 ) − β̂t,s+1
γp
s=t
L OG -L INEARIZED P HILLIPS C URVE
The role of financial frictions
π̂t
#
"
∞
X
ω(η − 1)
= −
µ̂t + Et
χδ̃ s−t+1 µ̂s+1 + βEt [π̂t+1 ]
γp
s=t
"
#
∞
X
1
s−t+1
χδ̃
+ [η − ω(η − 1)] Et
(ξˆt − ξˆs+1 ) − β̂t,s+1
γp
s=t
D EMAND S HOCK D URING THE F INANCIAL C RISIS
Homogeneous firms with nominal rigidities
(a) Output
(b) Hours worked
(c) Inflation
%
(d) Real wage
%
pps.
%
0.2
0.0
0.0
-0.5
0.2
0.1
0.1
0.0
-0.5
0.0
w/o FF
w/ FF
-0.1
-1.0
-1.0
-0.1
-0.2
-1.5
0
5
10
15
20
0
(e) Markup
5
10
15
20
-1.5
(f) Value of internal funds
pps.
0
5
10
15
20
-0.3
(g) Value of marginal sales
pps.
0
5
10
15
%
0.4
20
-0.2
(h) Interest rate
pps.
0.1
1.5
2.0
1.5
1.0
0.2
0.0
1.0
0.5
0.5
0.0
-0.1
0.0
0.0
-0.5
0
5
10
15
20
-0.2
0
5
10
15
20
0
5
10
15
20
-0.5
0
5
10
15
20
Financial crisis: ϕt = ϕ̄ = 0.5 (external finance premium = 20%)
-0.2
F INANCIAL S HOCK
Homogeneous firms with nominal rigidities
(b) Hours worked
(a) Output
(c) Inflation
%
(d) Real wage
%
pps.
%
1.5
w/ deep habit
w/o deep habit
0.0
0.0
-0.2
-0.2
-0.4
-0.4
-0.6
-0.6
0.2
0.0
1.0
-0.2
0.5
-0.4
0.0
0
5
10
15
20
-0.8
(e) Markup
0
5
10
15
20
-0.8
(f) Value of internal funds
pps.
0
5
10
15
20
0
(g) Value of marginal sales
pps.
15
5
0.8
4
4
0.6
3
3
2
2
1
1
0
20
0.6
0.4
0.2
0
0.0
0.0
0
5
10
15
20
0
5
10
15
20
-1
0
5
10
15
20
-0.6
pps.
5
0.2
10
%
1.0
0.4
5
(h) Interest rate
-1
0
Financial shock: ϕt = 0.3 → 0.375 (AR(1) dynamics)
5
10
15
20
H ETEROGENEOUS F IRMS
Two sectors that differ in operating efficiency: φ1 6= φ2
◮
Equal fixed measures of firms in each sector.
◮
Symmetric equilibrium within each sector.
Case I: φ1 = 0.8φ2 and φ2 = 0.3
◮
Financially more fragile economy with limited heterogeneity.
Case II: φ1 = 0 and φ2 = 0.3
◮
Financially more robust economy with greater heterogeneity.
Financial shock: ϕt = 0.3 → 0.375 (AR(1) dynamics)
“P RICE WAR ” IN R ESPONSE TO A F INANCIAL S HOCK
Heterogeneous firms with nominal rigidities
(a) Relative prices
(b) Output
%
%
0.6
0.5
Financially strong
0.4
0.0
Financially weak
0.2
-0.5
0.0
Financially weak
-1.0
-0.2
Financially strong
-1.5
-0.4
-0.6
0
10
20
30
40
Case I: φ1 = 0.8φ, φ2 = 0.3
Case II: φ1 = 0, φ2 = 0.3
-2.0
0
10
20
30
40
“P RICE WAR ” IN R ESPONSE TO A F INANCIAL S HOCK
Heterogeneous firms with nominal rigidities
(a) Relative prices
(b) Output
%
%
0.6
0.5
Financially strong
0.4
0.0
Financially weak
0.2
-0.5
0.0
Financially weak
-1.0
-0.2
Financially strong
-1.5
-0.4
Aggregate
Aggregate
-0.6
0
10
20
30
40
Case I: φ1 = 0.8φ, φ2 = 0.3
Case II: φ1 = 0, φ2 = 0.3
-2.0
0
10
20
30
40
“P RICE WAR ” IN R ESPONSE TO A F INANCIAL S HOCK
Heterogeneous firms with nominal rigidities
(a) Relative prices
(b) Output
%
%
0.5
0.6
Financially strong
0.4
Financially weak
0.0
0.2
-0.5
Financially weak
0.0
-1.0
-0.2
-1.5
-0.4
Financially strong
Aggregate
Aggregate
-0.6
0
10
20
30
40
Case I: φ1 = 0.8φ, φ2 = 0.3
Case II: φ1 = 0, φ2 = 0.3
-2.0
0
10
20
30
40
“P RICE WAR ” IN R ESPONSE TO A F INANCIAL S HOCK
Heterogeneous firms with nominal rigidities
(a) Relative prices
(b) Output
%
%
0.5
0.6
Financially strong
0.4
Financially weak
0.0
0.2
-0.5
Financially weak
0.0
-1.0
-0.2
Aggregate
Aggregate
0
10
20
-1.5
-0.4
Financially strong
30
40
Case I: φ1 = 0.8φ, φ2 = 0.3
Case II: φ1 = 0, φ2 = 0.3
Aggregate
Aggregate
-0.6
-2.0
0
10
20
30
40
C ONCLUSION
Internal liquidity positions and customer markets importantly influenced
firms’ price-setting behavior during the 2007–09 crisis:
◮
Liquidity unconstrained firms decreased prices, while liquidity
constrained firms increased prices.
◮
Differences in price-setting behavior concentrated in nondurable goods
industries.
DSGE model with customer markets and financial frictions:
◮
Significant attenuation of inflation dynamics in response to demand and
financial shocks.
◮
◮
Severe downturn in response to temporary financial shocks.
Tradeoff regarding inflation vs. output stabilization in response to demand
and financial shocks.
◮
“Paradox of financial strength” with heterogeneous firms.