Transcript Presentazione di PowerPoint

Regional economic modelling and
CSF 2000-2006 policy analysis:
the Italian experience in using
MOMEZ for the Mezzogiorno
Guido Pellegrini - University of Bologna and Ministry of
Economy and Finance
Aline Pennisi - Public Investment Evaluation Unit,
Department for Development Policies (DPS)
UVAL Open days – 4th July 2006
The econometric model for the Mezzogiorno
(MOMEZ)
The econometric model for the
Mezzogiorno (MOMEZ) was built by the
Department for Development and
Cohesion Policies, as an instrument for
the evaluation of the impact of the MDP
(Mezzogiorno Development Plan, now CSF
2000-2006) on growth of Southern
regions in Italy
SEMINAR ON MACROECONOMIC MODELLING AND CSF POLICY ANALYSIS
Outline of the presentation
MOMEZ and the CFS 2000-2006 logic
Theoretical and econometric issues in
setting a regional model for evaluation
The structure of MOMEZ and the main
differences with a typical UE Hermin-type
models
Some results about multipliers and
elasticities of MOMEZ
General evaluation uses of the model
Building different policy scenarios
Some results and recent examples
MOMEZ and the
CFS 2000-2006 logic
MOMEZ is strictly linked to the CSF logic
• The growth mechanism: the basic idea of the CSF
(and of the model) is that growth comes from the
effects of actions on the economic and social context,
that lead to an increase in productivity. This induces the
accumulation of human and physical capital, and
increases output.
• The policy effect channels: Policy effects are
captured by changes of public expenditure in
infrastructure and state aids. Moreover, changes in
expectations and in the social and economic
environment induced by policies can be captured by
some specific indicators that are named in CSF and in
MOMEZ as “break-through” variables.
SEMINAR ON MACROECONOMIC MODELLING AND CSF POLICY ANALYSIS
Theoretical and
econometric issues in
setting a regional model
for evaluation
Theoretical and econometric issues in
model setting
Estimating a regional econometric model for
evaluation in a dualistic economy requires
to deal with several issues:
• Modelling a dualistic country
• How to model the growth mechanism
• How to capture the externalities induced
by the policy
SEMINAR ON MACROECONOMIC MODELLING AND CSF POLICY ANALYSIS
Growth mechanism in a dualistic country /1
• Mezzogiorno
reflects
an
incomplete
transformation of the Italian economy, with the
prevalence of small and low-productivity firms in
manufacturing, agriculture and trade and the
specialization in traditional sectors.
• A dual economy is a not well integrated
structure, where both a developed area and a
more traditional and developing one co-exist in
the same country.
• The Northern Italy and the Mezzogiorno are a
good example of a dualistic system.
Growth mechanism in a dualistic country /2
There are 3 main sources of growth that a developing
area can experience within a dualistic environment:
• the “neighbourhood” effect: the more advanced
neighbourhood can play a conducive role in the
development of the Southern regions through the
spread of technology and the locomotive demand
effect. Even a “close” area can be affected by the
international demand through
the role of the
neighbourhood area
•
Policy actions can accelerate development
offsetting market failures and improving growth
transmission channels
•
Spillover effects may affect agents' expectations
modifying the links between regions.
Growth mechanism in a dualistic country /3
In MOMEZ:
• We adopted a biregional approach, where
Mezzogiorno and the Center-North area are
modelized in a similar way.
• The model is inserted in
a competitive
market
environment,
where
large
intertwining and several spillovers across
markets and regions exist.
• We explicitly consider the presence of policy
effects and externalities.
Modelling the growth mechanism /1
• A theoretically –founded growth model should
include an accumulation equation related to
different production factors.
• A developing area suffers not only shortage of
physical capital, but also a mismatch in human
capital: the stronger the complementarity
between physical and human capital, the wider
the gap. This can explain why the relevant
production constraints come from the
supply side rather than the demand side.
Modelling the growth mechanism /2
In the medium-long run the crucial bottleneck in
growth is the endowment of production
factors.
The choice is therefore in favour of a supply-oriented
model, with a central role played by accumulation
mechanisms.
This is coherent with the desired impacts of regional
policies, that are mainly in areas such as
infrastructure, human capital, and productivity.
This choice implies that the setting is only slightly
affected by short-term demand shocks, and
suggests that the model can be meant for longterm analysis.
Modelling the growth mechanism /3
.
Another big issue: the econometric way of
modelling the economic take off of a group
of adjacent regions.
•If a local economy takes off, the behaviour and
the expectations of agents in the area change
radically with respect to the past.
•Therefore, the parameters of a model describing
the previous behaviour of the economy will also
change.
•Technically, here the Lucas critique clearly
applies. And data on structural changes are
usually not available.
Modelling the growth mechanism /4
.
There are two ways to tackle this problem:
•to follow Lucas critique and identify the “deep
parameters” modelling expectations in an explicit
way
•To use a calibration technique which relies on
parameters estimated in a different context.
In our analysis we use both of these
approaches.
How to model the policy impact /1
The task of modelling
channels is not easy.
policy
transmission
The mere inclusion of proxy variables for
different kinds of public expenditure cannot
capture the multiplicative effects and the
interactions coming from changes in both the
expectations
and
the
socio-economic
environment of the taking off economy.
How to model the policy impact /2
One solution is to choose variables for expectations
and changes in social and economic environment
due to policy actions.
We decided to use some specific indicators that
were named “break-trough” variables.
These variables were chosen because in those areas
in which signals of remarkable local development
occurred they improved significantly (at the end of
the previous regional policies in Italy)
They give a sign that a break-trough with the past is
occurring.
The 13 key breakthrough variables and their indicators
Break-through variable
Indicator
Export capacity
export/GDP
Economic
independence
They
are
chosen throughneta import/GDP
critical analysis of the economic
time (days
per person) spent
touristic
situation and of the structural
characteristics
ofinthe
southern
accomodation/population
Attraction of tourism
economy
Rate of capital accumulation
gross fixed investment/GDP
Attraction of foreign direct
investment
Labour market participation
direct foreign investment in Italy /direct foreign
investment in the EU
Participation rate
High productivity export
Non regular employed/total employed (labour
units)
exports of goods of increasing world demand /
tot. Exports
Innovation
Public and private exp. for R&D / GDP
Supply of regular employment
employed in business services/ total employed in
services
Expansion of business services
If
the
expected
“break”
in the Mezzogiorno actually
Expansion of personal qualified people age 14 or more participating in volontary
be caught
byetc../
thepopulation
evolution
these
services takes place, it willwork,
associations,
of 14 of
years
difference of the rate of interest between Centrevariables.
Development of financial markets North and Mezzogiorno
Law enforcement and social
cohesion
index of violent crime / 1000 inhabitants
Main features of the model
and differences with
Hermin-type models
Main features of the model
•
The model captures the main growth mechanisms
of the developing areas of a dualistic economy and
selects the principal transmission channels of
public policies.
•
it is focused on the supply side; disequilibria
between demand and resources are solved by
import-export from the Centre-North (not through
price equilibria)
• it is disaggregated into 4 sectors (agriculture,
industry, services, public administration)
• It consists of 38 behavioural equations and 60
identities (about 90 variables in time-series since
1980, of which 40 exogenous)
•
different multipliers for capital public expenditure
related to infrastructure and state aid
Exogenous variables
Break-through
variables
(low and high
hypotheses)
Pu.Ad. capital
expenditure
(QFU)
GDP
Labour market,
wages and incomes
TFP
Public Invest.
sud
State aids
sud
K (kapital)
Private
investment
Capital
utilization
Centro-Nord
(sub-model)
consumption
B.P.
International
variables
export
Public
investm.
centro-nord
import
Differences with Hermin-type models
The
logic is similar (supply-side model; effects
through TFP) but:
- our model is disaggregated only into four
sectors (Hermin more)
- Hermin-type models have a price endogenous
cycle. We do not have price circle.
- in Hermin-type elasticities to policy variaibles
are fixed values, calibrated outside the model’s
frame; in our case they are estimated within the
model
- in our model the whole strategy underlying the
CSF and all public policies for the development of
least favoured areas (EU and national) are taken
into account
- basically our model is designed to evaluate our
policy strategy, Hermin a standard strategy
WHAT THE MODEL IS NOT
It is not a tool designed for forecasting, because it
is not oriented to short term analysis
It is not a tool built up to analyze the Southern
economy, because not all possible transmission
channels are taken into account.
MULTIPLIER ANALYSIS
The true test of the model is in the evaluation of
the impact of several shocks.
We use a standard procedure, shocking some
exogenous variables and comparing the results
with a baseline solution. This procedure allows us
to empirically calculate the multipliers and to show
the long run properties of the model.
This is basically a test of theoretical and empirical
consistence of the model.
.
Shock on world demand
Percent Deviation
pil aggregato p. 95 sud (somma VA settoriali+ sbi+altro)
.16
ula totali sud
.14
.12
..
.12
.10
.08
.08
.06
.04
.04
.02
.00
2001
2002
2003
2004
2005
2006
2007
2008
.00
2001
import nette sud
.30
-.1
.25
-.2
.20
-.3
.15
-.4
.10
-.5
.05
2002
2003
2004
2005
2006
2003
2004
2005
2006
2007
2008
inv tot tutti i settori pub e priv
.0
-.6
2001
2002
2007
2008
.00
2001
2002
2003
2004
2005
2006
2007
2008
Shock on public investment
Percent Deviation
pil aggregato p. 95 sud (somma VA settoriali+ sbi+altro)
.10
..
ula totali sud
.12
.10
.08
.08
.06
.06
.04
.04
.02
.00
2001
.02
2002
2003
2004
2005
2006
2007
2008
.00
2001
import nette sud
2002
2003
2004
2005
2006
2007
2008
inv tot tutti i settori pub e priv
.24
.5
.20
.4
.16
.3
.12
.2
.08
.1
.04
.00
2001
2002
2003
2004
2005
2006
2007
2008
.0
2001
2002
2003
2004
2005
2006
2007
2008
General evaluation uses of
the model
The model’s main evaluation tasks
 to evaluate the coherence of the programme
with its objectives given the macroeconomic and
other sustainability constraints (is the strategy
compatible with the stability pact, demographic trends,
international demand, etc..?)
 to identify the transmission channel of the
implementation of the development plan (starting
from public investments which other economic variables are
affected and to what extent? Is the strategy that is going to be
implemented affecting the right variables?)
 to give an estimation of the impact of the plan
on GDP and employment of the Objective 1
Regions : global target. (what is the growth rate of the
area that is expected as a consequence of the overall
intervention? what conditions must be met to guarantee the
targeted growth rates?)
MOMEZ is integrated within the framework
of the other models used at DPS
Regional public
accounts forecasting model
National develop.
Fund monitoring
system
European funds
monitoring
system
Regional
public accounts
Public capital
expenditure
frame (1996-2009)
Data bases
Macro-economic model
(MOMEZ)
Short and medium term territorial
growth forecast model
Data bases
21/07/2015
Official documents,
Economic and Financial Government Plan
Building different policy
scenarios
Preliminary idea on how the model works
Evaluation question:
….what growth rate is expected for the
Mezzogiorno area as a consequence of the
overall intervention? etc…
Exogenous variables:
Public policy variables
(infrastructure/state-aid expenditure)
Breakthrough variables
International variables
MACROECONOMIC
MODEL
Different growth scenarios
representing different
hypotheses on the
exogenous variables
Four basic scenarios used for evaluation
purposes
1. Baseline / counterfactual scenario (no
increase foreseen in public expenditure
according to specific a investment plan)
2. Planned public capital investment only
(“historical trend” for breakthrough
variables – no externalities)
3. Planned public capital investment with low
externalities (“optimistic trend” for
breakthrough variables)
4. Planned public capital investment with
high externalities (“more optimistic trend”
for breakthrough variables)
Examples and recent
results
Verifying consistency with global objectives
….. Achieve for the Mezzogiorno a
significantly higher growth rate than in the
EU (average EU-15)
5,0
4,5
4,0
3,5
EU-15 GDP
3,0
2,5
2,0
1,5
1,0
0,5
0,0
1999
2000
historic values
2001
2002
2003
Today - no externalities
2004
2005
2006
Today - low ext.
2007
2008
Today - high ext.
2009
2010
EU-15 GDP
Mezzogiorno growth rate (annual % change) – various scenarios
Monitoring the global objectives: current
scenarios compared to those of the ex-ante
evaluation
6,0
5,0
4,0
3,0
2,0
1,0
0,0
1999
ex-ante progr
historic values
Why
2007 ? 2008
What
today's
prog happened ?
ex-ante low ext.
ex-ante high ext.
today low ext.
2000
2001
2002
2003
2004
2005
2006
today high ext.
Mezzogiorno growth rate (GDP annual % change) – various scenarios
25,0
Today’s scenario versus the ex-ante
scenario: quantifying “responsibilities”
20,64
20,0
5,1 due to reduced
expenditure
15,0
5,2 due to context
10,0
10,4
5,0
0,0
Today's scenario
Due to reduced expenditure
2000-2008
Due to change of exogenous var. (context)
Due to the change in model specification
Ex-ante conditions with today's model
Mezzogiorno growth rate (GDP % change 2000-2008)
Changes in the international context
EU-15 GDP growth rate
4,0
3,5
3,0
2,5
2,0
1,5
1,0
0,5
World imports growth rate
0,0
1998
1999
2000
2001
2002
ex-ante
2003
2004
intermediate
2005
today
2006
2007
2008
14,0
12,0
10,0
8,0
6,0
4,0
2,0
0,0
1998
1999
2000
2001
2002
ex-ante
2003
2004
intermediate
2005
today
2006
2007
2008
Changes in total public capital
expenditure
25,0
24,0
observed expenditure in 2006
observed expenditure in 2003
23,0
billion euro
22,0
21,0
20,0
19,0
18,0
17,0
16,0
15,0
1999
2000
2001
Ex-ante (1999)
2002
2003
2004
2005
Intermediate eval. (2003)
2006
2007
2008
Today (march 2006)
Public capital expenditure in the Mezzogiorno (real values)
2009
Changes in breakthrough variables / 1:
“attracting tourism” – comparing ex-ante,
intermediate and today’s expectations
7,00
Attracting tourism
6,00
5,00
4,00
3,00
2,00
1,00
0,00
1995
1996
1997
1998
ex ante low
1999
2000
ex ante high
2001
2002
2003
intermediate low
2004
2005
2006
intermediate high
2007
2008
today low
2009
2010
today high
2011
2012
Changes in breakthrough variables / 2:
“law enforcement” – comparing ex-ante,
intermediate and today’s expectations
18,00
17,00
Crime rate index
16,00
15,00
14,00
13,00
12,00
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
basse exante
alte exante
basse int
alte int
basse attuale
alte attuale
Evaluation of impact respect to
counterfactual situation
5,0
4,0
3,0
2,0
1,0
0,0
-1,0
1999
2000
2001
2002
historic values
Today - high ext.
2003
2004
2005
2006
Today - no externalities
Counterfactual scenario
2007
2008
2009
Today - low ext.
Mezzogiorno growth rate (annual % change) – various scenarios
2010