Topic 6 Evaluation HO

download report

Transcript Topic 6 Evaluation HO

We Have Developed a
Theoretical Model Capable
of Explaining All the
Stylised Facts
So Now We Have a Theoretical
Model Capable of Explaining All the
Stylised Facts
In fairness, some results depend on some
effects being greater or less than others
(e.g. hours worked, interest rates etc)
It is an empirical question whether this is
actually the case.
So Now We Have a Theoretical
Model Capable of Explaining All the
Stylised Facts
So we need to build a model economy like
Chapters, 2, 4, 5 and 12,
with realistic parameters for the
production function, tastes for work and
leisure, and investment function etc
and check whether such an economy can
replicate the observed stylised facts.
Evaluation and Assessment
• Readings (on reading list):
• Ryan & Mullineux- The Ups and Downs of
Modern Business Cycle Theory in
Reflections on the Development of Modern
Macroeconomics, Snowden & Vane eds
For a contrary view see paper by Dixon in
same volume
Evaluation and Assessment
• Additional readings which I have placed on the
J drive:
• Ryan - Business Cycle Theory: The Real
Business Cycle in Encyclopaedia of
Macroeconomics (Forthcoming 2002)
• Ryan – Business Cycle Theory: The Stylised
Facts Real Business Cycle in Encyclopaedia of
Macroeconomics (Forthcoming 2002)
Origins of the Real Business
Cycle Theory
1. Theoretical Failures & Add ons
2. Lucas Critique
3. Rational Expectations
Origins of the Real Business
Cycle Theory
1. Theoretical Failures & Add ons
r
LM
IS
y
1970’s No
Supply side
in ISLM
Model
Demand
Side model
Only
r
LM
IS
y
P1
P0
AD
y1
y0
Then expected
that if y went
down P would fall.
So when observed
P rising and y
falling needed
new construct
r
P1
P0
LM
Added on Supply
IS
which when we
had rational
expectations had
to be vertical
y
AS1
AS0
AD
y1
y0
No real
Microeconomic
Foundations
Origins of the Real Business
Cycle
2. Lucas Critique
Can’t use reduced forms to evaluate policy
e.g. recall macro last year y=c+i+G-T
If c=a+byd where yd is disposable income
and yd=(1-t)y
Then dy
1
dG

1  b(1  t )
Origins of the Real Business
Cycle
dy
1

dG 1  b(1  t )
2. Lucas Critique
If change Government policy (e.g. G or T)
then may change marginal propensity to
consume, b.
b is a derived parameter, and can’t rely on it
Origins of the Real Business
Cycle
dy
1

dG 1  b(1  t )
2. Lucas Critique
Need to use basic theory such as utility and
production functions not derived
consumption functions such as
c=a+byd
Origins of the Real Business Cycle
3. Rational Expectations
Up to 1970’s used to believe that Monetary
Policy was very important
BUT Rational Expectations showed that
people could be expected to figure out the
broad trust of monetary policy and thus
ability if the government to ‘surprise’ them
was limited
So Monetary Policy is ineffective
Origins of the Real Business Cycle
3. Rational Expectations
Monetary Policy is ineffective
If still believe that Money is important then
need a proper model of how money works
in the economy
– not a crappy LM curve
SO again need to return to a more basic model
Origins of the Real Business
Cycle Theory
So have three reasons to return to basic
micro modelling
• Theoretical Failures & Add ons
• Lucas Critique
• Rational Expectations
Origins of the Real Business
Cycle
Finally, a comment on shocks:
What was causing a business cycle?
How were these shocks propagated?
Origins of the Real Business
Cycle
Old belief: Supply Shocks small
Demand and Monetary Shocks more important.
But if government monetary policy is anticipated
then monetary shocks must be less important
And Demand side shocks must be due to MPKe
that is, supply side shocks that effect D
SO need to look more closely at Supply side shocks!
Real Business Cycle Model
Original Kydland & Prescott (1982)
Economica
Utility Function
Production Function

U t  Ct L t

y t  K t 1N t
1
1 
CALIBRATING the Model
Use MICRO econometric studies to decide on
(the CES) form of the functions and
appropriate values for  and 
Utility Function
Production Function

U t  Ct L t

y t  K t 1N t
1
1 
WHY COBB-DOUGLAS
Because Shares of K, N and C & L relatively
constant over time: Remember 201 Test 1
Utility Function
Production Function
U t  Ct L t
.25
yt  K Nt
.33
t 1
.75
.66
Recall we said we could think
of production function as –
y
Good year
Normal
Bad year
N
yt 
y
Zt
.33
t 1
K Nt
Good year
.66
But need to
include a
Normal
shock Zt
Bad year
for good
and bad
years
N
1
3
t 1
y t  Zt K N t
y
Zt
N
2
3
So now we are ready for our
computer simulation model
Stick the Utility function, the production
function and the investment Function in
the computer, switch it on and …….
y
yG
y0
yB
N
NB N0 NG
And the computer churns out:
•
•
•
•
•
•
N0, NB, NG,………………
y0, yB, yG,………………
c0, cB, cG,………………
i0, iB, iG,………………
p0, pB, pG,………………
Etc etc for each state Zt
• And now we are almost there:
• All we have to do now is look at
how each variable fluctuates
compared with real economy
• So how do we do?
Evaluation
• How does the Model Do?
• Answer: Simulation is reasonably
CLOSE except for variability of hours
worked
• But how can we be sure this is a good
model?
• Some controversy!!
Evaluation
1. Are the facts a fact?
• Vary across time
– E.g. Prices between 1929+ 1939
– and 1945 - now
• Vary across country
– broad results the same some difference
• How are these facts determined?
– How do we fit the Long Run Trend Line
Evaluation
2. How Close is CLOSE?
• No measure of Closeness
– Not like Confidence Interval in
Econometrics
• Can perform sensitivity tests
– Use slightly different parameters for
utility and production functions
– New econometric techniques developed
to cope with problem
– Generalised Method of Moments
Evaluation
3. The need for labour Market
modifications
a) 40 Hour Week
b) Allowing the intensity with which
capital is worked to vary
c) recognising that work effort in one
period might influence desire to work in
the next
All these things improve performance of
the model
•
Results
of
Exercise
Topic
7
Bottom Line:
• Changes in w/p over the cycle will not
induce much variation in hours worked
• However, changes in overtime rates
will (pure substitution effect)
• Main effect will be on marginal
individual making the work/no-work
decision
• In particular for higher w/p people are
more likely to make the effort to be in
employment
What are the implications of this for our
Labour Demand & Supply diagrams?
W/P
lD
lS
W/P
lD
lS
W/P
lD
lS
lS’
ll
lu
W/P
lD
lS
lS’
ll
lu
W/P
lD
lS
lS’
ll
lu
W/P
lD
lS
lS’
ll
lu
One other problem with labour
market!
• The Dunlop-Tarshis Test
Policy implications:
• Macroeconomic fluctuations are the
result of optimal responses to
economic shocks.
• If responses are Optimal then
Government can’t do any better
• Short-run Government policy
responses distorts these optimal
responses
• Particularly true of Monetary Policy
Policy implications:
• Doesn’t men no role for government
• In fact would claim that the policy
implications are very Keynesian in one
sense
• The structure of the economy is all
important
• So getting ‘balance and structure’ of
government policy is absolutely vital
for the long-run growth path of the
economy
Policy implications:
• Not about Right Wing Policy Agenda
• About Methodology
• Must have Macroeconomics with proper
micro-foundations
• Better Labour market sector – but properly
modelled ( Next part of Course)
• Next development-a Proper modelled
monetary Sector