The International Propagation of News Shocks Paul Beaudry, Martial Dupaigne & Franck Portier University of British Columbia & Universit´ e de Toulouse SED Meeting, 06.28-30.2007 Prague
1
1. Motivation • News shocks:
data : Beaudry & Portier [2006, Aer; 2005, Jjie], Haertel & Lucke [2007],
models : [Beaudry & Portier [2004, Jme; 2007, Jet], Christiano, Rostagno & Motto [2005], Jaimovich & Rebelo [2006], Den Haan & Kaltenbrunner [2006], Beaudry, Portier & Collard [2007]
• Technological News Shocks: Short run demand shock, Long run supply shock • A source of international fluctuations? Small Open Economy: Jaimovich & Rebelo [2007]; Two-country economies: this paper 2
1.1. Business cycle comovements • Y , C, I, H are positively correlated with each other within developed countries, at business cycle frequencies National Business Cycle (NBC) • Y , C, I, H are pairwise positively correlated among developed countries, at business cycle frequencies International Business Cycle (IBC) • Which combination(s) of impulses and propagation mechanisms can help understand these business cycle co-movements?
3
1.2. The effects of technological shocks • The international RBC literature faces huge difficulties to account for international comovements. • If countries experience different technology shocks, mobile inputs reallocate to the most productive economy, and the returns to immobile inputs lower. Extremely correlated technology shocks are required to match the observed correlations of inputs. • “Demand” shocks might help. Wen [2006, Jecd]
4
1.3. The nature of technological shocks • The usual assumption is that technology shocks are surprises. • Beaudry & Portier [2006, Aer] show that (permanent) technology improvements diffuse slowly over time, and are forecastable to a large extent. • In the short–run, these news shock stimulate the demand for investment goods, and might not trigger reallocation.
5
Outline of the Talk
1. Motivation
2. The Propagation of News Shocks : Facts
3. NBC and IBC in a canonical model
4. NBC and IBC in an extended model
6
2. The Propagation of News Shocks: Facts 2.1. Conditional moments • If technological change diffuses slowly over time, ‘forward’ variables may react faster than usual indicators of technology. • We identify news shock using TFP (corrected for utilization) and stock market capitalization (SP )
7
• BP 2006:
∆T F Pi,t ∆SPi,t
!
= A(L)
ε1,t ε2,t
!
h
= I+
P+∞ k A L k k=0
i
ε1,t ε2,t
!
.
- the news shock ε2,t has no impact on TFP in country i;
∆T F Pi,t ˜ • Here ∆SPi,t = A(L) Xj,t
ε1,t h i P+∞ k ˜k L ε2,t = I + k=0 A ε3,t
ε1,t ε2,t ε3,t
Ak
˜k = with A
× ×
0 0 . ×
- the news shock ε2,t has no impact on TFP in country i;
- the third shock ε3,t has no impact on TFP and stock prices in country i. 8
2.2. US news shocks and their propagation • (Corrected TFP, SP) VECM with 5 lags. • The US news shock has a significant long–run effect on US TFP and explains a large share of the forecast error. • It has almost no impact on US TFP during the first five years ⇒ this is not a TFP surprise. Response to a news shock, USA 1
0.7 CTFP
0.8
0.6
0.6
0.5
0.4
0.4
0.2
0.3
0
0.2
−0.2
0.1
−0.4 0
5
10
15
20
25
30
0 0
CTFP
50
100
150
200
9
Response to a news shock, USA 1.4
3.5
0.8
C
I
N
1.2
3
1
2.5
0.8
2
0.4
0.6
1.5
0.2
0.4
1
0.2
0.5
0
0
10
20
1.5
0
0.6
0
0
10
20
1.4
−0.2
0
10
20
0.05
Y
C+I+X−M
(X−M)/Y
1.2 1
0
1 0.8
0.5
−0.05 0.6 0.4
0
−0.1
0.2 −0.5
0
10
20
0
0
10
20
−0.15
0
10
20
10
• A news shock triggers an expansion in Canada as well as in the US. Response of Canadian aggregates to a news on US TFP 1
2.5
1
C
I
N
0.8
2
0.8
0.6
1.5
0.6
0.4
1
0.4
0.2
0.5
0.2
0
0
0
−0.2
−0.5
−0.2
−0.4
0
10
20
1.5
−1
0
10
20
1.5
−0.4
0
20
0.6
Y
C+I+X−M
1
10
(X−M)/Y 0.4
1
0.2 0.5
0.5 0
0
−0.5
0
0
10
20
−0.5
−0.2
0
10
20
−0.4
0
10
20
11
2.3. German news shocks and their propagation • German data are from Haertel & Lucke [2006]. (Corrected TFP, SP) VECM with 2 lags. • The permanent improvement in TFP takes place after 4 years. Response to a news shock, Germany 1
0.5 CTFP
0.8
CTFP
0.4
0.6
0.3
0.4 0.2
0.2
0
0.1 −0.2 −0.4 0
5
10
15
20
25
30
0 0
50
100
150
200
12
Response to a news shock, Germany 2
2
0.6
C
I
N
1.5
1.5
0.4
1
1
0.2
0.5
0.5
0
0
0
−0.2
−0.5
0
10
20
2
−0.5
0
10
20
2
−0.4
0
20
0.3
Y
C+I+X−M
1.5
10
(X−M)/Y 0.2
1.5
0.1 1
1
0
0.5
0.5
−0.1
0
0
−0.2
−0.5
0
10
20
−0.5
−0.3 0
10
20
−0.4
0
10
20
13
Response of Austrian aggregates to a News on German TFP 3
2 C
0.4 I
2
N 0.2
1
1
0 0
0 −1 0
10
20
1.5
−1 0
−0.2 10
2 Y
10
20
0.4 C+I+X−M
1
(X−M)/Y 0.2
1
0.5
0 0
0 −0.5 0
−0.4 20 0
10
20
−1 0
−0.2 10
−0.4 20 0
10
20
14
Response of French aggregates to a News on German TFP 1.5
5
1.2
C
I
N 1
4 1
0.8
3
0.6 0.5
2 0.4 1
0.2
0 0 −0.5
0
10
20
1.2
−1
0 0
10
20
2
−0.2
C+I+X−M
1
10
20
0.2
Y
(X−M)/Y 0.1
1.5
0.8
0
0.6
1
−0.1
0.4
0.5
−0.2
0.2
−0.3 0
0 −0.2
0
0
10
20
−0.5
−0.4 0
10
20
−0.5
0
10
20
15
Response of Bristish aggregates to a News on German TFP 2
4
1.2
C
I
1.5
N 1
3
0.8 1
2
0.6
0.5
1
0.4
0
0
0.2
−0.5
0
10
20
1.2
−1
0 0
10
20
1.5
−0.2
0
10
20
0.1
Y
C+I+X−M
(X−M)/Y
1
0 1
0.8
−0.1
0.6
−0.2 0.5
0.4
−0.3
0.2
−0.4
0
0 −0.2
−0.5 0
10
20
−0.5
0
10
20
−0.6
0
10
20
16
Response of Italian aggregates to a News on German TFP 1.5
3
0.3
C
I
1
N
2.5
0.2
2
0.1
1.5
0
1
−0.1
0.5
−0.2
0
−0.3
0.5
0
−0.5
0
10
20
1.5
−0.5
0
10
20
1.5
−0.4
0
20
0.3
Y
C+I+X−M
1
10
(X−M)/Y 0.2
1
0.1 0.5
0.5 0
0
−0.5
0
0
10
20
−0.5
−0.1
0
10
20
−0.2
0
10
20
17
2.4. What have we learned? • Conditional on news to future TFP, main macro aggregates display strong comovements across countries. • We now try to account for these findings.
18
3. NBC and IBC in a canonical model • Here we show that in a canonical model, news shocks is a IBC driving force
19
3.1. The model • A 2-country, 1-good economy. The economy is hit by technology shocks θA,t and θB,t. Capital quantity and location are predetermined. o
n
in order to • Choose Cj,t, Hj,t, Ij,t, Kj,t+1 j=A,B max E0
+∞ X
h i t β U CA,t, 1 − HA,t + U CB,t, 1 − HB,t
t=0
subject to
KA,t+1 ≤ (1 − δ ) KA,t + IA,t KB,t+1 ≤ (1− δ ) KB,t + IB,t CA,t + CB,t + IA,t + IB,t ≤ F KA,t, HA,t; θA,t + F KB,t, HB,t; θB,t | | {z } {z }
YA,t
KA,0 = KB,0 given
YB,t
.
• We make the further simplifying assumption that preferences are separable in consumption and leisure (U12 = 0). 20
3.2. Some Propositions • Some propositions can be proved, that show the respective role of local/global/surprises/news in creating NBC and IBC.
21
Result 1 In response to global surprises (θA,t = θB,t ∀t), equilibrium allocations are symmetrical. The model displays IBC. Under functional and parameters restrictions, the model also displays NBC.
22
World Technological Surprise 1.5 ΘA ΘB
1 0.5 0
0
2
4
6
8
10
0.5
15 CA CB
0.45
10
0.4 0.35
5
0
2
4
6
8
10
1.5
0
0
2
4
6
8
10
1 YA YB
1
HA HB
0.5
0.5 0
IA IB
0
0
2
4
6
8
10
−0.5
0
2
4
6
8
10
23
Result 2 If technology shocks are local and surprises (dθA,t > 0, dθB,t = 0 for some t), then hours worked are not perfectly correlated across countries. For realistic settings, hours and investments are negatively correlated. There is therefore no IBC and no NBC in the foreign country.
24
Local Technological Surprise 1.5 Θ A Θ
1
B
0.5 0
0
2
4
6
8
10
0.3
4000 CA CB
IA IB
2000
0.25
0 −2000
0.2
0
2
4
6
8
10
100
−4000
2
4
6
8
10
100 YA YB
50
0
−50
−50 0
2
4
6
8
HA HB
50
0
−100
0
10
−100
0
2
4
6
8
10
25
Result 3 If technology shocks are announced/forecastable N periods in advance, then allocations are symmetrical in the N − 1 first periods of the interim period, for both world and local news ⇒ IBC. In the interim period, consumption and hours always move in opposite directions ⇒ no NBC.
26
(World) Technological News 1.5 ΘA ΘB
1 0.5 0
0
2
4
6
8
10
0.08
−2 CA CB
0.06
IA IB
−4 −6
0.04 0.02
−8 0
2
4
6
8
0
−10
2
4
6
8
−0.2 YA YB
−0.2
−0.6
−0.6
−0.8 0
2
4
6
HA HB
−0.4
−0.4
−0.8
0
8
−1
0
2
4
6
8
27
(Local) Technological News 1.5 ΘA ΘB
1 0.5 0
0
2
4
6
8
10
0.05
4000 CA CB
0.04
2000
0.03
0
0.02
−2000
0.01
0
2
4
6
8
0
−4000
0
2
4
6
8
−0.2 YA YB
−0.2
−0.4
−0.6
−0.5 0
2
4
6
HA HB
−0.3
−0.4
−0.8
IA IB
8
−0.6
0
2
4
6
8
28
4. An Extended Model • We build on Beaudry & Portier [2004, jme] “Pigou model” • Building blocks are :
1. Two sectors in each countries (Consumption and Investment (structures))
2. Capital and Labor are complementary in the consumption good sector
3. There are static gains to trade (Armington aggregators for consumption and investment goods)
+ Investment is produced with labor only, with DRS 29
4.1. Model o
n
˜j,t, Ij,t, Kj,t+1 in order to Choose Cj,t, Hj,t, H j=A,B max E0
+∞ X
h i t ˜A,t + ln CB,t − χ HB,t + H ˜B,t β ln CA,t − χ HA,t + H
t=0
s.t.
KA,t+1 XA,t ZA,t
≤ (1 − δ ) KA,t + IA,t αX f ˜ H ≤ Θ A,t A,t ≤
CA,t ≤
i1 ϕ ν ν a ΘA,tHA,t + KA,t h i1 νC νC bZAA,t + (1 − b)ZBA,t νC h i1 νI νI bXAA,t + (1 − b)XBA,t νI
h
IA,t ≤ idem country B KA,0 = KB,0 given.
30
4.2. Result: Local Technological News to ΘA 1.5
ΘA ΘB
%
1 0.5 0
0
2
4
6
8
10 6
CA CB
0.5
0
0
2
4
2
6
8
0
10
0
2
4
6
8
10
4
6
8
10
1
YA YB
0.5
0
2
%
%
1
0
IA IB
4 %
%
1
4
6
8
10
HA HB
0.5
0
0
2
31
4.3. To Sum Up • News shocks are observed to create NBC and IBC • One can design almost standard models to account for this • In progress: Reproduce VARs conditional responses with simulated data • In progress: Check for unconditional moments
32
