Short overview
Slides to Lecture 1 of Introductory Dynamic
Macroeconomics
1. ’Statics and dynamics’
Concepts and definitions. Lecture 1.
2. Dynamic analysis
Models and methods. Lecture 2 and 3.
3. Wage-price dynamics. Lecture 4 and 5.
Ragnar Nymoen
University of Oslo, Department of Economics
4. Macro dynamics. Lecture 6-14.
• Review of demand side (closed economy); Role of asset markets
• Stabilization policy, rules versus discretion
January 16, 2006
• Open economy AS-AD model. Short-and long run.
• Foreign exchange market. Regime dependency.
• Current monetary policy
1
1
See more detailed plan at http://folk.uio.no/rnymoen/ECON3410 v06 index.html
which is the workpage of the course.
2
‘Statics and dynamics’, an introduction
Economic agents typically take time to adjust their behaviour to changes in
circumstances–because of habits, learning, norms and other institutions (for
example annual wage setting).
Instantaneous adjustment is the exception in economics. Adjustments lags
represent the rule.
At the aggregate (macro) level: a shock typically affects the economy several
periods after the it first occurred–the effects of a shock are dynamic.
Expectations:
Backward-looking expectations are another explanation of lags
Forward-looking expectations sometimes creates a particular form of dynamics:
large adjustments first (“over-shooting”), then more gradual adjustment.
3
4
Time lags as a premise for decision making.
Norges Bank [The Norwegian Central Bank] is typical of many central banks’
view:
By the way, before it was changed in the summer of 2004, the same passage
on Norges Banks internet pages read:
“Monetary policy influences the economy with long and variable lags.
Norges Bank sets the interest rate with a view to stabilizing inflation
at the target within a reasonable time horizon, normally 1-3 years”
“A substantial share of the effects on inflation of an interest rate
change will occur within two years. Two years is therefore a reasonable
time horizon for achieving the inflation target of 2 12 per cent.”
Hence policy decisions–meaning interest rate setting–is based on Norges
Banks beliefs about the dynamic nature of the monetary transmission mechanism.
Taken at face value, this shows that Norges Bank’s belief/model of the transmission mechanism changed in the summer of 2004. Later in the cause we will
be able to assess the likely impact of such a change on policy decisions!
In economics, beliefs means models, implicit or explicit.
Hence, the statement illustrates two theses: policy is model based, and policy
models are dynamic.
5
6
A static model
A static demand schedule, by definition:
Definition: ‘statics and dynamics’
Dynamic analysis is a relatively new invention.
It all began with Frisch who worked intensively with the foundations of the
discipline he dubbed macrodynamics in 1929 and the first years if the 1930s.
His operational definition of dynamics is:
A dynamic theory or model is made up of relationships between variables that
refer to different time periods. Conversely, when all the variables included in
the theory refer to the same time period (or, more generally, the model is
conceptualized without time as an entity), the system of relationships is static.
Xt = aPt + b + εd,t,
with a < 0 and b > as parameters.
The variables X and P,and εd (denoting a random demand shock) are all
provided with time subscript t.
t might represent for example a year (the time period is annual); or a quarter
(the period is quarterly); or month (the period is monthly).
In theoretical models we usually set t = 0, 1, 2, ....T . Where t = 0 is referred
to as the initial period and T the terminal period. Sometimes though, the
terminal period is not specified, so we write t = 0, 1, 2, .....
In empirical models, t refers to calendar time periods: years, quarter, months
for example.
7
8
If we supplement the demand equation with a static supply equation, we obtain
the static marked equilibrium model
Xt = a Pt + b + εd,t, demand, and
εs,t
Supply shocks
0.01
0.00
<0
Xt = c Pt + d + εs,t,
>0
supply.
determining the endogenous variables Xt and Pt for known values of the exogenous variables εd,t and εs,t (and fixed and known values of the 4 parameters
a − d).
We assume that εd,t and εs,t random variables. Their role is to represent
shocks, or in Frischean terminology, impulses to the system. We do not need
to be specific about the distribution of εd,t and εs,t: but if it might be helpful
to fix ideas to think of them as normally distributed with zero mean and a
constant variances.
-0.01
0
75
20
40
60
80
100
120
140
time
Density
50
25
-0.0175 -0.0150 -0.0125 -0.0100 -0.0075 -0.0050 -0.0025 0.0000 0.0025 0.0050 0.0075 0.0100 0.0125 0.0150 0.0175
Figure 1: Example of normally distributed supply shocks.
9
Pt
10
Nature of exogenous variables
Later in the course we will also introduce exogenous variables with a non-zero
average and with a large deterministic component.
Demand curve
(average position)
P1
Supply curve (average position)
C
B
D
Such variables are exogenous economic explanatory variables.
P0
A
They may be determined on the world marked, or controlled by the government
for example.
But for the time being, in order to discriminate clearly between static and
dynamic models, it is convenient to only use purely random exogenous variables.
X1
X0
Xt
Figure 2: A static marked equilibirum model. IDM fig 1.1.
11
12
Static marked equilibrium model
Characteristics of a static model
Static marked equilibrium model
1.00
0.03
market price
Dynamic multipliers
net demand shock
0.02
The whole effect of a shock is contained in the equilibrium values of P and X
in the period of the shock.
0.75
0.01
0.50
0.00
0.25
-0.01
There is no spill-over effects of a shock in period t = 1 to period 2, 3, ....
10
20
30
40
50
Dynamic marked equilibrium model
Impulses in period 1 are not propagated to later periods
0
1.0
5
10
15
20
Dynamic marked equilibrium model
0.050
Dynamic multipliers
market price
net demand shock
0.5
0.025
• The time series of Pt (and Xt) are perfect mirror images of the shocks εd,t
and εst. Fig 3 a).
0.0
0.000
-0.5
-0.025
• The sequence of dynamic multipliers, for example ∂Pt/∂εs,1 (t = 1, 2, 3...)
are zero, expect for ∂Pt/∂εs,1. Fig 3 b)
13
10
20
30
40
50
0
5
10
15
20
Figure 3: Static model and dynamic (cobweb) model. IDM fig 1.2.
14
Assuming that t = 0 represents the initial situation with εs,0 = 0. According
to the dynamic model, supply in period t = 1 is given by
A dynamic model (cobweb)
Xt = a Pt + b + εd,t, demand, and
<0
Xt = c Pt−1 + d + εs,t,
>0
supply.
The only change is in the supply equation, where Pt−1 replaces Pt.
By Frisch’s definition the supply equation is now dynamic, since the relationship
is in terms of variables from different time periods.
This is a seemingly trivial change in specification. But it affects both the
economic interpretation and the propagation of shocks fundamentally.
Interpretation: In some markets supply is fixed in the short-run. No matter
how high or low the price is in the current period, the supply of the good is
‘frozen’ by decisions of the past. Classic example: agricultural products such
as pork and wheat.
15
X1 = cP0 + d.
Supply in period 1 is a function of P0 which is pre-determined from history,
and not the price in period 1. The supply in period 1 (short-run supply) is thus
completely inelastic with respect to the price, P1.
Pt−1 is therefore called a pre-determined variable.
The parameter c is thus not a short-run slope coefficient in the way is was in the
static model, instead it gives the response of supply to a lasting change in price.
We define c as the long-run slope coefficient, for the stationary relationship
X̄ = cP̄ + d,
where are stationary values of supply and price (Xt = Xt−1 = X̄ and Pt =
Pt−1 = P̄ )
16
Pt
“Short-run” and “long-run” are not absolute, but relative to the model. With
Xt = c Pt−1 + d + εs,t,
Long-run supply curve
supply.
>0
we only have to wait one period to see the full response of a price change in
period t on supply. Xt does not change, but Xt+1 changes with derivative
coefficient c.
Demand curve
(average position)
B
P1
D
Supply is elastic with a one period lag. In other modes this will be different,
typically the response is more sluggish.
A
P0
C
P2
But even this simple case of one-period supply lag has a large impact on the
behaviour of the equilibrium values of price and quantity, compared to the static
model.
To see how, assume that t = 0 is the initial period, and also a stationary
situation with X0 = X̄ and P0 = P̄ , and εd,0 = εs,0 = 0.
X0
X2
Xt
Figure 4: The cobweb model. IDM fig 1.3
Then consider a demand shock in period 1: εd,1 > 0.
17
Static marked equilibrium model
18
Characteristics of dynamic models
Static marked equilibrium model
1.00
0.03
market price
Dynamic multipliers
net demand shock
0.02
The whole effect of a shock is no longer contained in the equilibrium values of
P and X in the period of the shock.
0.75
0.01
0.50
0.00
There are spill-over effects of a shock in period t = 1 to period 2, 3, ....
0.25
-0.01
10
20
30
40
50
Dynamic marked equilibrium model
0
1.0
5
10
15
20
Dynamic marked equilibrium model
0.050
Impulses in period 1 are propagated to later periods
Dynamic multipliers
market price
net demand shock
• The time series of Pt (and Xt) are not mirror images of the shocks εd,t
and εst.
0.5
0.025
0.0
0.000
-0.5
-0.025
10
20
30
40
50
0
5
10
15
20
• The sequence of dynamic multipliers, for example ∂Pt/∂εd,1 (t = 1, 2, 3...)
are generally non-zero, but may approach zero for large values of t, if the
dynamics is stable.
Figure 5: Copy of figure 3: Static model and dynamic (cobweb) model.
• Dynamics is a system property: though only a single relationship in system
is dynamic, the response of the whole system is affected (in general)
19
20
Dynamic market equilibrium model, habit formation.
Another dynamic model (‘habit formation’)
1.00
0.050
market price
Dynamic market equilibrium model, habit formation.
net demand shock
dynamic multipliers
0.75
0.025
Xt = aPt + bXt−1 + εd,t, demand and
Xt = c0Pt + c1Pt−1 + d + εs,t.supply
The demand function is now a dynamic equation. The parameter b is no
longer a constant intercept, but instead measures by how much an increase in
Xt−1 shifts the demand curve. This can be rationalized by habit formation for
example, in which case we may set 0 < b < 1.
0.50
0.000
0.25
-0.025
10
20
30
40
Dynamic market equilibrium model, cobweb.
50
1.0
0
5
10
15
Dynamic market equilibrium model, cobweb.
0.050
20
Dynamic multipliers
market price
net demand shock
0.5
0.025
0.0
In any given period, Xt−1 is determined from history and cannot be changed.
Hence in this model there are two pre-determined variables: Pt−1 and Xt−1.
0.000
-0.5
-0.025
10
20
30
40
0
50
5
10
15
The supply equation is a generalization of the cobweb model, with supply
showing some simultaneous response to a price change (coefficient c0).
Figure 6: Two dynamic models. IDM fig 1.4.
21
22
The cobweb model and habit model are both dynamic.
20
The short-run and the long-run
In the static model, the full effect of a shock on the endogenous variables were
reached in the same time period as the shock.
But properties are different:
• Cobweb amplifies shocks, habit dampens shocks (compare panel c) and
a))
• Dynamic multiplier: changing signs (‘pork cycles’), and geometrically declining (habit formation)
The telling difference is that in the two dynamic models, the endogenous variables’ response to shocks were not instantaneous, and the system took several
periods to adjust to a shock in period t.
The exact shapes of the responses, which we dubbed dynamic multipliers above,
depend on the detailed dynamic model specification.
Illustrates that dynamic models (even without nonlinearity and forward-looking
expectations) are flexible, can be used to represent many form of behaviour of
economic variables.
We draw from this that static models as a framework for analysis, is best suited
when the speed of adjustment of the variables are so fast that we can ignore
that ‘actually’ there is some time delay between the impulse (or shock) and
the response.
23
24
Frisch:
“Hence it is clear that the static model world is best suited to the
type of phenomena whose mobility (speed of reaction) is in fact so
great that the fact that the transition from one situation to another
takes a certain amount of time can be discarded. If mobility is for
some reason diminished, making it necessary to take into account the
speed of reaction, one has crossed into the realm of dynamic theory.”
The choice between a static and a dynamic analysis will therefore depend on
what we judge to be realistic or typical of the phenomenon which is the subject of our study: very high speed of reaction to impulses, or more moderate
response time.
Paradoxically, phenomena which in an everyday meaning of the word are really
dynamic, with lots of volatility, as for example stock market prices, can be
analyzed scientifically using a static framework;
At least as a first approximation, which brings out that the choice between
static and dynamic analysis may also be relative to the level of ambition of our
study, and to the time and other resources available for modelling
25
26
However, apart from resource constraints and a need to simplify out of necessity,
the scales are tilted in the favour of dynamic models in macroeconomics: Frisch:
For example: The true intellectual rationale for the IS-LM model is that we
think it captures some really important and dominant aspects of the macroeconomy over a period of 1-2 years for example.
We do not deny that a fiscal policy stimulus for example have other effects
beyond that short horizon. The static IS-LM model by definition cannot help
us in the modelling of those effects, which therefore remain beyond the scope
and ambition of the IS-LM based analysis.
One of the goals of our course is to extend the analysis of monetary and fiscal
policy into the realm of dynamic theory.
“In real life both inertia and friction act as a brake on speed of
reaction”.
Frisch also noted that
“The static theory’s assumption regarding an infinitely great speed
of reaction contains one of the most important sources of discrepancy
between theory and experience.”
Therefore Frisch anticipated the increased use of dynamics in economics–it
would increase the degree of realism and scope of macroeconomic analysis.
27
28
Short-run and long-run models
However, there is also another, quite different, interpretation of static models
which will figure prominently in this book.
It is that static equations express what the dynamic model would correspond to
in a counterfactual situation where no shocks, impulses or changes in incentives
occurred.
Recall the habit formation model. Figure 6 a) above (fig 1.4 in IDM) shows
the full solution for Pt. Sometimes, when we are unable to derive the solution
of the dynamic model, we will still be able to answer two important questions:
1. What are the short-run effect of a change in an exogenous variable? And
2. What are the long-run effects of the shocks.
As we become accustomed to dynamic analysis, we will refer to this correspondence by saying that static relationships can represent the stationary state (or
steady state) of a dynamic model.
The technique we will use is based on the distinction between the short-run
model given by
Xt = aPt + bXt−1 + εd,t, (demand)
It is custom to use the terms stationary (or steady state) equation interchangeably with the term long-run equation.
Xt = c0Pt + c1Pt−1 + d + εs,t, (supply),
taking the predetermined variables Xt−1 and Pt−1 as exogenous, and the longrun model which is defined for the stationary situation of Xt = Xt−1 = X̄, Pt =
Pt−1 = P̄ and εd,t = ε̄d, εs,t = ε̄s.
29
30
P
Short-run supply curve
1
a
P̄ +
ε̄d, long-run demand, and
1−b
1−b
X̄ = (c0 + c1)P̄ + ε̄s, long-run supply
Long-run supply curve
X̄ =
The long-run model applies to a hypothetical stationary situation where there
are no new shocks, and all past shocks have worked their way through (and
“out of”) the system.
Graphically we can represent the short-run and long-run models in one diagram,
using lines with different slopes to illustrate the difference between the shortand long-run.
C
B
A
Short-run demand curve
Long-run demand curve
X
We can then analyze the short-run, or impact, effect of a shock (to demand
for example), as well as the long-run effect.
Figure 7: Graphical analysis of a demand shock in the habit formation model
31
32
Stock and flow variables
Starting from a stock variable like Pt, a flow variable results from obtaining
the change, hence
Dynamic models often include both flow and stock variables.
• Flow: in unit of (for example) million kroner per year
• Stock: in units of (for example) million of kroner at a particular period in
time (for example start or end of the year).
In this course an important class of stock variables will be price indices. For
example Pt may represent the value of the Norwegian CPI in period t (a month,
a quarter or a year).
As you know, the values of P will be index numbers. The number 100 (often
1 is used instead) refers to the base period of the index. If Pt > 100 it means
that relative to the base period, prices are higher in period t.
CPI in Norway
• zt ≈ yt by the properties of the (natural) logarithmic function, see for
example the appendix of IDM, if in doubt.
An typical empirical trait of stock variables are that they change gradually, as
a result of finite growth rates.
1500
Sometimes however, stock variables jump from one value in period t to quite
another in period t + 1.
1000
200
1.5
• yt × 100 is inflation in percentage points. In this course we often stick to
the rate formulation (hence, we omit the scaling by 100)
CPI in the UK
400
1760
are examples of flow variables. Note that:
34
33
600
the (absolute) change
xt = Pt − Pt−1,
Pt − Pt−1
yt =
,
the relative change, and
Pt−1
zt = ln Pt − ln Pt−1 the approximate relative change
1780
1800
1820
1760
0.4
Inflation in Norway
1.0
1780
1800
1820
Inflation in the UK
Empirically, the rate of change then becomes very large. Norwegian “price
history” at the breakdown of the union with Denmark is an example.
0.2
In economics, when stock variables change gradually, we need explicit dynamic
models to account for their evolution.
0.5
0.0
0.0
-0.5
-0.2
1760
1780
1800
1820
1760
1780
1800
1820
Sometimes though, stock variables can be treated theoretically as if the are
jump-variables. An example of such a theory in this course is the portfolio
model of the foreign exchange market. That model is static.
Figure 8: Consumer price indices (stock variables), and their rate of change
(flow). Norway and UK
35
36
75
The Norwegian current account
Billion kroner
50
As noted economic dynamics often arise from the combination of flow and
stock variables.
25
0
For example, the dynamic behaviour of debt (a stock) is linked to the value of
the current account (flow) in the following way
1980
1985
1990
1995
2000
1990
1995
2000
Norwegian net foreign debt
debt = − current account + last periods debt + corrections.
For example: If there is a primary account surplus for some time (and ignoring
corrections for simplicity), this will lead to a gradual reduction of debt–or an
increase in the nation’s net wealth. Conversely, a consistent current account
deficit raises a nation’s debt.
Billion kroner
0
-250
-500
-750
1980
1985
Figure 9: The Norwegian current account (upper panel) and net foreign deb
(lower panel)t. Quarterly data 1980(1)-2003(4)
37
38
The textbook consumption function, i.e., the relationship between real private
consumption expenditure (C) and real households’ disposable income (IN C)
is an example of a static equation
Using quarterly data for Norway, for the period 1967(1)-2002(4), we obtain by
using the method of least squares:
Ct = f (IN Ct), f 0 > 0.
ln Ĉt = 0.02 + 0.99 ln IN Ct
An empirical example
(1)
Two examples of specified functional forms for the static consumption function:
Ct = β0 + β1IN Ct + et,
(linear)
ln Ct = β0 + β1 ln IN Ct + et, (log-linear)
(2)
If this is unfamiliar, read in IDM about the properties of these two functional
forms. For example for the interpretation of β1.
where the “hat” in Ĉt is used to symbolize the fitted value. Next, use (2) and
(3) to define the residual êt:
êt = ln Ct − ln Ĉt,
which is the empirical counterpart to the static model disturbance et.
Textbooks usually omit the term et in equation (2), but as we have seen above,
when we want a more precise interpretation of macroeconomic equations it is
important to include a disturbance term, in the same way as the shock terms
εd,t and εs,t above.
39
(3)
40
(4)
The dynamic consumption function:
12.0
ln Ct = β0 + β1 ln IN Ct + β2 ln IN Ct−1 + α ln Ct−1 + εt
(5)
is an example of a so called autoregressive distributed lag model, ADL. The
estimated version is:
11.8
11.6
ln C t
ln Ĉt = 0.04 + 0.13 ln IN Ct + 0.08 ln IN Ct−1 + 0.79 ln Ct−1
(6)
Compare residuals ε̂t and êt to judge which model is best (see graph).
11.4
Which explanatory variables contribute most to the improved fit?
11.2
ln Ct−1 itself! Illustrates that the dynamic framework is important.
11.0
11.0
Figure 10: The estimated static consumption function.
The low estimated income elasticities (0.130 and 0.08) reflect that Norwegian
households have a low propensity to consume income rises that may turn out to
be transitoy. The next lecture show that these results imply a much propensity
to consume a permanent change in income.
41
42
11.1
11.2
11.3
11.4
11.5
11.6
11.7
11.8
11.9
12.0
ln INCt
The road immediately ahead
0.15
Residuals of (2.7)
Residuals of (2.4)
Next week we start a more systematic introduction to the modelling tools of
dynamics, see IDM Ch 2.
0.10
0.05
Present a general framework for dynamic single equation models, called the
autoregressive distributed lag model, ADL model.
0.00
−0.05
Using that model, the concept of the dynamic multiplier, encountered several
times already can be made precise.
−0.10
−0.15
The dynamic multiplier is a key concept in this course
−0.20
−0.25
1965
1970
1975
1980
1985
1990
1995
2000
Figure 11: Residuals of the two estimated consumptions functions (3), and (6),
The ADL model also defines a typology of dynamic equations, as special cases,
each with their distinct features.
The analysis is extended to simple dynamic systems-of-equations.
Then we turn to wage-and-price setting, which is an essential part of the supply
side of modern macroeconomic model.
43
44