Tools for Monetary Policy with DSGE
Models
g3 – case studies
Jaromír Tonner
Brno, autumn, 2011
Outline
• Expert judgments
• Case studies
• Sensitivity analysis
2
Expert judgments
•
•
•
•
Measurement errors
Seasonal adjustment
Expert judgments on filter
Expert judgment on forecast
3
Measurement errors
• ME reflect our priors concerning data reliability.
• ME brings some problems in distinguishing between
structural shock and measurement error.
• Even in case of ME, a significant portion of information
can be used by the model.
• Another problem is that filtered vars need not match
exactly raw data, so then ...
• ...we investigate factors for that discrepancy...what are
models or data deficiencies.
4
Measurement errors
5
Seasonal adjustment
6
Expert judgment on filter
7
Expert judgment on forecast
• All forecasts are judgemental forecast (calibration of the model,
filtering setup, trajectories of structural shocks), but
• we may impose judgements on the development of a particular
variable by endogenizing structural shocks innovations, but....
• the question is... what shock or set of shocks to choose and
whether these shocks should be treated as anticipated or
unanticipated...in which periods
• A special case represents explaining of a current development of a
given variable by future innovations...these must be treated as
anticipated by all agents in the economy...
• A solution is not unique, we can choose the set of shocks that is
the most likely...
8
Case studies outline
•
•
•
•
•
Nominal wages
Increase of credit premium
Foreign trade fall and openness
Car scrapping subsidies
After-crisis steady states growth rates
9
Case studies of nominal wages
•
•
•
•
•
Real variables slump, nominal wages sticky
It implies inflation pressures
Model in reduced form
Find reason
Expert judgment
10
Case studies of nominal wages
Real GDP growth, (%, y-o-y)
Nominal wages growth, (%, y-o-y)
8
12
6
10
4
8
2
0
6
-2
4
-4
-6
I/06
I/07
I/08
I/09
I/10
2
I/06
I/07
I/08
I/09
I/10
11
Case studies of nominal wages
Median of gross month salary (Kč/month)
1. half of 2008
1. half of 2009
Firms
21 330
21 309
Employees
22 064
21 837
Fired
18 478
New
18 115
Source: prezentation Spáčil, Kasal: Projevy krize v datech ISPV, www.trexima.cz
• Firms could keep the same wages only if fired were
employees with low salary and...
12
Case studies of nominal wages
Number of employees (thousands)
1. half of 2008
1. half of 2009
Firms
1 707
1 619
Employees
1 383
1 409
Fired
323
New
210
• 1/3 less new employees than fired
Source: prezentation Spáčil, Kasal: Projevy krize v datech ISPV, www.trexima.cz
13
Case studies of nominal wages
Hours worked monthly
1. half of 2008
1. half of 2009
Firms
153.6
149.9
Employees
154.8
149.3
Fired
148.6
New
154.1
• Fired have the lowest hours worked
Source: prezentation Spáčil, Kasal: Projevy krize v datech ISPV, www.trexima.cz
14
Case studies of nominal wages
Hours of sickness monthly
1. half of 2008
1. half of 2009
Firms
6.1
5.0
Employees
5.1
5.0
Fired
10.3
New
4.8
• Fired have the highest hours of sickness.
Source: prezentation Spáčil, Kasal: Projevy krize v datech ISPV, www.trexima.cz
15
Case studies of nominal wages
• Who lost job?
•
•
•
•
•
Low wages employees
Double sickness absence
Low hours worked
Low qualified
Basic education
• It increases average wages, however we need
fundamental wages.
• It requires expert judgment…how much?
16
Case studies of nominal wages
Q-o-Q wage growth (seasonally adjusted)
10
8
% (annualized)
6
4
2
0
-2
2006
Wage
Wage
Wage
Wage
Wage
Wage
2006.5
growth
growth
growth
growth
growth
growth
2007
(wages)
(compen)
per employee (wages)
per employee (compen)
per hour (wages)
per hour (compen)
2007.5
2008
2008.5
2009
2009.5
2010
17
Case studies of nominal wages
Nominal wages growth (%, y-o-y)
12
observed
adjusted
10
8
6
4
2
I/00
I/01
I/02
I/03
I/04
I/05
I/06
I/07
I/08
I/09
I/10
18
Case studies of nominal wages
3M PRIBOR (%, p.a.)
Inflation (%, y-o-y)
8
orig wages
6
baseline
4
orig wages
4
baseline
3
2
2
0
I/06
I/08
1
I/06
I/10
GDP (%, y-o-y)
I/07
I/08
I/09
I/11
Nom. exchange rate CZ K/EUR
10
30
orig wages
baseline
5
0
-5
I/06
I/10
orig wages
baseline
28
26
I/07
I/08
I/09
I/10
I/11
24
I/06
I/07
I/08
I/09
I/10
I/11
19
Case studies of credit premium
• Observation of an increase of credit premium (interbank
3M PRIBOR rates did not follow changes of the 2W
CNB’s policy rate)
• Associated with uncertainty, releasing worse and worse
economic data (at the beginning of 2009) etc.
• g3 shows us a trajectory of the interbank rate but does
not incorporate policy rate (2W REPO rate)
• the setting of 2W REPO rate is based on the model
consistent trajectory of 3M PRIBOR interbank rate and a
credit premium
20
Case studies of credit premium
• Solution outside the model ! necessity of assumption
about a premium to recognize required trajectory of the
policy rate
• Assumption that the premium in the Czech economy
corresponds to a European Area premium
• Approximation using monthly data:
Cred premt =
3mPribort−(2wRepot+2wRepot+1+2wRepot+2)/3
21
Case studies of credit premium
22
Case studies of trade fall
• Observation of deep and equivalent slumps of export and
import volumes
• The Czech economy case: large import intensity of
exports (not only value added is traded ...)
• Because slumps of volumes implies a fall in re-exports,
the decrease of GDP is smaller
• Explanation: cut-off of some logistics centres with
relatively low added value
23
Case studies of trade fall
• g3 is well-suited for this situation and provides a tool for a
solution
• g3 incorporates the openness technology to capture the
different (and significant in reality) growth rates of trade
volumes with respect to output growth (re-exports)
24
Case studies of trade fall
• Initial conditions (model filtration): we observed a
decrease of the openness technology
• Part of volume slumps can be attributed to a decrease of
openness technology
• On the other hand, the near-term-forecast (NTF)
reassesses the nowcast of trade volumes in order to have
deeper slump of exports than imports (see the figure in
the introduction) -> a decrease of output
25
Case studies of car scrapping
subsidies
• Car scrapping subsidies in some European countries
raise demand for Czech cars
• Car industry is very significant in the Czech Republic
(Skoda, TCPA - Toyota Peugeot Citroen Automobile,
Hyundai, many suppliers)
• Car scrapping significant mostly for lower class cars (the
majority of Czech production) ! we observe an increase of
car production, and hence exports
• But subsidies in 2009, strong uncertainty for 2010
26
Case studies of car scrapping
subsidies
• Foreign demand modelled exogenously
• Expert judgement for foreign demand: significant increase
in 2009, but a decrease in 2010 with the termination of
subsidies
• People will exploit subsidies before the termination with
subsequent fall in demand for cars
• Reference paper: Adda, J. and Cooper R. (2000):
Balladurette and Juppette: A Discrete Analysis of
Scrapping Subsidies. The Journal of Political Economy
108:4.
27
Case studies of after crisis SS
• Relatively high steady-state growth rates before the crisis
• Uncertainty with the after-crisis growth (in the short- as
well in the long-run)
28
Case studies of after crisis SS
• We keep the same steady-state to have historical
filtrations unchanged
• But for the forecast the steady-state growth rates of
technologies are decreased...
• ...(assumption about the gradual recovery)
• We simulate various scenarios with (lower growth rates in
middle-run and long-run, lower wages, negatively shocks
to technologies etc.)
29
Scenario analysis
• Scenario vs. Fan charts (graphs with confidence intervals)
• Scenario analysis is constructed to capture uncertainty of the
produced forecast.
• Scenario analysis also serves the purpose of gaining better
intuition.
• Scenarios may differ not only in alternative paths of exogenous
variables but also whether and what variables are anticipated or
unanticipated.
• Our decomposition tools are:
• decomposition of alternative scenarios into factors,
• analysis of sources of a difference between two successive forecast,
• dynamics decomposition of a forecast w.r.t the steady state.
30
Scenario analysis
•
•
•
•
Nominal wages and productivity
Foreign demand
Energy prices
Exchange rate
31
Sensitivity to nominal wages and
productivity
•
•
•
•
•
Four experiments
Observation of higher wages is shock
Observation of higher wages due to higher productivity
Expert judgment of lower wages on forecast is shock
Expert judgment of lower wages on forecast due to lower
productivity
32
Sensitivity to nominal wages and
productivity
Inflation CPI (y-o-y, %)
3M PRIBOR (%)
3
baseline
6
higher wages - shock
2.5
5
2
4
lower wages - shock
3
1.5
2
1
0.5
I/10
higher wages and prod.
lower wages and prod.
1
II
III
IV
I/11
II
III
IV
I/07 III
I/08 III
I/09 III
I/10 III
I/11 III
33
Sensitivity to nominal wages and
productivity
Nominal exchange rate CZK/EUR
Real GDP growth (y-o-y, %)
baseline
26
4
higher wages - shock
3
higher wages and prod.
lower wages and prod.
25.8
25.6
lower wages - shock
2
25.4
1
25.2
25
I/10
II
III
IV
I/11
II
III
IV
0
I/10
II
III
IV
I/11
II
III
IV
34
Sensitivity to nominal wages and
productivity
Tempo růstu reálné spotřeby (y-o-y, %)
Nominal wages growth (y-o-y, %)
6
baseline
2
higher wages - shock
5
1
4
0
3
-1
2
-2
1
I/10
-3
I/10
II
III
IV
I/11
II
III
IV
higher wages and prod.
lower wages and prod.
lower wages - shock
II
III
IV
I/11
II
III
IV
35
Sensitivity to foreign demand
PRIBOR 3M (%, p.a.)
Real GDP growth (y-o-y, %)
4.5
8
4
6
3.5
4
3
2
2.5
0
2
I/06
III
I/07
III
I/08
III
I/09
III
I/10
III
Nominal exchange rate (CZK/EUR)
-2
I/06
10
28
9
27
8
26
7
25
6
24
5
III
I/07
III
I/08
III
I/09
III
I/10
I/07
III
I/08
III
I/09
III
I/10
III
Nominal wages growth (y-o-y, %)
29
23
I/06
III
III
4
I/06
III
I/07
III
I/08
III
I/09
III
I/10
III
36
Sensitivity to energy prices
Sensitivity scenario of oil prices in deviations from baseline
Oil prices growth (y-o-y, %)
Gas prices growth (y-o-y, %)
Foreign inflation PPI (y-o-y, %)
Regulated prices inflation (y-o-y, %)
Inflation CPI (y-o-y, %)
3M PRIBOR (%)
Nominal exchange rate (CZK/EUR)
Real GDP growth (y-o-y, %)
III/09 IV/09
6.6 18.8
4.2
2.2
0.6
0.6
0.3
0.1
0.0
0.0
0.3
0.2
0.0 -0.1
0.1
0.2
I/10
22.8
7.4
0.6
0.7
0.1
0.4
-0.2
0.0
II/10
16.1
18.3
0.5
1.0
0.2
0.3
-0.2
-0.2
III/10 IV/10
1.7 -3.5
18.7 10.2
0.2
0.2
0.9
1.1
0.2
0.3
0.1
0.2
-0.1 -0.1
-0.3 -0.1
I/11
-3.0
3.0
-0.1
0.4
0.1
0.0
-0.1
-0.2
37
Sensitivity to exchange rate
Sensitivity scenario in deviations from baseline
Inflation CPI (y-o-y, %)
3M PRIBOR (%)
HDP (y-o-y, %)
Nominal exchange rate (CZK/EUR)
I/10
0.0
0.0
-0.2
-0.8
II/10 III/10 IV/10
0.0 -0.1 -0.2
-0.3 -0.3 -0.3
-0.2 -0.1 0.0
-0.2 -0.1 0.0
I/11
-0.2
-0.1
0.2
-0.1
II/11 III/11
-0.1 0.0
-0.1 0.1
0.3 0.2
-0.1 -0.1
38