Transcript An updated post-Keynesian alternative to the New consensus on

A post-Keynesian alternative to the New
consensus on monetary policy
Marc Lavoie
University of Ottawa
What is the New consensus?
• New Keynesian “consensus-type model”,
where policy reaction functions are “an
essential part of the macroeconomic
system”.
• Also called, the New Keynesian Synthesis
• Their formal version is the DSGE model
• J.B. Taylor, Blinder, D. Romer, Woodford,
Meyer.
The New Consensus
• “There is substantial evidence demonstrating that
there is no long-run trade-off between the level of
inflation and the level of unused resources in the
economy – whether measured by the
unemployment rate, the capacity utilization rate,
or the deviation of real GDP from potential GDP.
Monetary policy is thus neutral in the long run.
An increase in money growth will have no longrun impact on the unemployment rate; it will only
result in increased inflation.” Taylor 1999
The new consensus: summary
• The new consensus is simply a variant of
monetarism;
• but without any causal role for money.
• It is monetarism without money.
New consensus model:
A simplified version
•
•
•
•
•
•
IS function:
g = g0 ! $r + ,1
Vertical Phillips curve
dB/dt = ((g ! gn) + ,2
Central bank reaction function:
r = r0 + "1(B ! BT) + "2(g ! gne)
• A similar model could be build with an output gap
or a rate of capacity utilization instead of g
How is the New consensus linked to
post-Keynesian theory?
• Positively
• Negatively
• “The main change is
• The New consensus
that it replaces the
reproduces accepted
assumption that the
dogma among
central bank targets the
neoclassical economists,
money supply with an
la pensée unique as the
assumption that it
French say.
follows a simple interest
rate rule” (Romer 2000:
154).
La pensée unique
• Expansionary fiscal policy only leads to
higher inflation rates and higher real interest
rates in the long run;
• More restrictive monetary policy only leads
to lower inflation rates in the long run.
How is the New consensus linked to
post-Keynesian theory ?
(2) About the reaction function
• Post-Keynesians
• There is an interest
rate reaction function
because it cannot be
otherwise. Money
supply targeting is
impossible in principle
and in practice.
• New Consensus
• There is an interest
rate reaction function
because interest rate
targeting is more
successful to dampen
shocks to money
demand than money
supply targeting.
How is the New consensus linked to
post-Keynesian theory ?
(3) About the IS function
• New Consensus (see Dullien)
• Post-Keynesians
• The IS function is based on
• The IS function is
aggregate supply. A higher
based on aggregate
real interest rate reduces
demand, with
present consumption, lowers
investment reacting
wages and real wages (prices
negatively to the real
are rigid, adjust with lags), and
interest rate.
hence reduces the supply of
labour and output.
New consensus model:
An alternative version (Setterfield)
•
•
•
•
•
•
IS function:
g = g0 ! $r + ,1
Vertical Phillips curve
dB/dt = ((g ! gn) + ,2
Central bank reaction function:
dr = "1(B ! BT) + "2(g ! gne)
Implicit to
the New Keynesian model
• A natural real interest rate
• r0 = rn = (g0 ! gn)/ $
• which implies (g ! gn) = $(r – rn)
• A natural growth rate, given by supply-side factors
• gn = a constant
Figure 5: The hidden consensus equation
B
0
gn
gn
Kaldor: The Scourge of Monetarism
Assuming that the behaviour of the `real' economy is
neutral with respect to monetary disturbances, why
should the elimination of inflation be such an
important objective as to be given 'over-riding
priority'? In what way is a community better off with
constant prices than with constantly rising (or falling)
prices? The answer evidently must be that … inflation
causes serious distortions and leads to a deterioration in
economic performance, etc. In that case, however, the
basic proposition that the `real' economy is impervious
to such disturbances is untenable. Pp. 41-42
2000 survey of AEA economists
• Question: Real GDP eventually returns
automatically to potential real GDP?
– Agree: 63 %
– Disagree: 37 %
– Journal of Economic Education, Fall 2003
2000 survey of AEA economists
• Question: There is a natural rate of
unemployment to which the economy tends
in the long run ?
– Agree: 68 %
– Disagree: 32 %
r
Figure 1: The graphical new consensus with
T)
r
=
r
+
"
(B
!
B
0
1
r
RF


MP
IS
B
B
g
B


AD
B
Start with the historically given rate of inflation
gn
g
Figure 2: New consensus with rising MP curve
r
r
RF
MP


IS
B
B
B

g

AD
B
gn
g
Figure 3A: Impact of a rise in effective demand:
There is a lag in inflation
r
r
r3
RF
r1
B
A

MP

IS
B
B
g
B
B3
B1


B
A
IA
AD
B
gn
g2
g
Figure 3: Impact of a rise in effective demand
r
r
r3
RF
r1
C
A

B


MP
IS
B
B
g
B
B3
B1


C

B
A
IA
AD
B
g3=gn g2
g
Figure 4A: Bringing back inflation to its target rate
r

RF4

r4
r3
D
C

r1

B


MP
A
IS2
RF1
IS1
B
B
B
B3
g
D

BT
B

A
g4
C

gn

B
g2
IA
AD2
AD1 AD4
g
Figure 4: Bringing back inflation to its target rate
D
r

RF4

r4
r3
D
C

r1

B


MP
A
IS2
RF1
IS1
B
B
B
B3
g
D

BT
B

A
g4
C

gn

B
g2
IA
AD2
AD1 AD4
g
Post-Keynesian alternatives
• (1): Reject the vertical Phillips curve and
replace it with a long-run downwardsloping Phillips curve (Setterfield) or a flat
Phillips curve (Kriesler-Lavoie)
• Or,
• (2) Endogenize the natural rate of growth gn
(Lavoie)
Similarity with PK critique of
natural rate of unemployment
• Hargreaves-Heap (Economic Journal 1980)
• Cottrell (JPKE, 1984-85)
• dUn /dt = N(U ! Un) + ,3
Post-Keynesian views
• “Disequilibrium adjustment paths can affect
equilibrium outcomes” (Colander, 1996:
60), leading to multiple equilibria and to
path-dependent equilibria.
• “The natural rate of growth is ultimately
endogenous to the demand-determined
actual rate of growth” (Setterfield, 2002: 5)
Post-Keynesian views II
• “Neoclassical growth economists on the one
hand, ... treat the rate of growth of the
labour force and labour productivity as
exogenous to the actual rate of growth....
• Economists in the Keynesian/postKeynesian tradition, ... maintain that growth
is primarily demand driven because labour
force and productivity growth respond to
demand growth” (León-Lesdema and
Thirlwall, 2002).
Post-Keynesian views III
•
“But at the same time technical progress is being
speeded up to keep up with accumulation. The rate of
technical progress is not a natural phenomenon that falls
like the gentle rain from heaven. When there is an
economic motive for raising output per man the
entrepreneurs seek out inventions and improvements. Even
more important than speeding up discoveries is the
speeding up of the rate at which innovations are diffused.
When entrepreneurs find themselves in a situation where
potential markets are expanding but labour hard to find,
they have every motive to increase productivity”
(Robinson 1956, p. 96).
Post-Keynesian views IV
• “The stronger the urge to expand … the greater are the
stresses and strains to which the economy becomes
exposed; and the greater are the incentives to overcome
physical limitations on production by the introduction of
new techniques. Technical progress is therefore likely to be
greatest in those societies where the desired rate of
expansion of productive capacity … tends to exceed most
the expansion of the labour force (which, as we have seen,
is itself stimulated, though only up to certain limits, by the
growth in production)” (Kaldor, 1960, p. 237).
Kaldor 1933 revisited
• ‘We shall call an equilibrium “determinate” or
“indeterminate” according as the final position is
independent of the route followed or not [path
dependence]; we shall call equilibrium “unique”
or “multiple” according as there is one, or more
than one, system of equilibrium prices
corresponding to a given set of data; and
finally,we shall speak of “definite” or “indefinite”
equilibria, according as the actual situation tends
to approximate of position of equilibrium or not’
[stability or instability].
2000 survey of AEA economists
• Question: Changes in aggregate demand
affect real GDP in the short run but not in
the long run ?
– Agree: 62 %
(potential growth is given)
– Disagree: 38 % (potential growth is affected
by short-run effective demand)
The alternative PK model
g = g0 ! $r + ,1
dB/dt = ((g ! gn) + ,2
r = r0 + "1(B ! BT) + "2(g ! gn)
r0 = rnT = (g0 ! gn)/ $
dgn /dt = N(g ! gn) + ,3
PK model becomes a system of two
differential equations
 dg / dt 
 dg / dt 
 n


  (1   2  )   1 

1   2



  (1   2  )   1 
1   2





  3   2  (d 1 / dt ) 
g  

1   2
g   

 n 
3


PK model becomes a system of
two differential equations
• The dynamics of this system are pretty
straightforward, because the determinant of
this system is zero;
• and because its trace, is always negative.
• The amended new consensus model
displays a continuum of equilibria. The
model is said to contain a zero root.
• As the long-run equilibrium is not
predetermined anymore, the steady-state
rate of accumulation now depends on
transitional dynamics, which cannot be
ignored: short-run events have a qualitative
impact on long-run equilibria. It is common
to speak of ‘path-dependence’ for such a
characteristic. It is possible to show that this
kind of model displays hysteresis in the
sense of a ‘permanent effect of a transitory
shock’ (Olivier 1999).
An example of hysteresis
• For instance, a temporary increase in the
rate of price inflation that would arise
independently of excess demand pressure
would have permanent effects on the natural
rate of growth and the natural real rate of
interest.
Another example
• Monetary policy now has real effects that
go beyond its impact on the inflation rate.
• Zero-inflation or low-inflation targeting has
a negative impact on the real economy,
bringing in high real rates of interest and
low real rates of growth.
Figure 6: Path-dependence, likely case
gn
dg = 0
dgn = 0
A*
gn
B

E


A
B*
gB gnB gnE gnA gA
g
Figure 8A: Reducing the inflation target: PK model
r
rB
rB *
r
RF2


B

C

rC
rE


E
IS
B
B
g
B
BE

B
BC
AD2

BT
B

E
C
AD

gB
gnC gnE
g
Figure 8: Reducing the inflation target: PK model
r
r
rB
rB *
RF3
RF2



B

B*

C

rC
rE


E
IS
B
B
g
B
BE


B
BC
BT
B


C
E
AD
B*
gB gnB gnC gnE g
Other visual interpretation of hysteresis
Slope is
gn = n + a
Log K
Log qfc
New slope
Actual path
Time
The New Consensus view
Log K
Log qfc
Slope is
gn = n + a
Actual path
Time
A less drastic interpretation of hysteresis, when the
central bank reacts to utilization rates, with
g = gn + µ (u  un)
Log K
Log qfc
Slopes are
gn = n + a
Actual path
Time
Other critiques: IS Curve
Reject the simple interest rate -investment
relation implied in the IS model:
• Central bank sets short term interest rate, but
what about the long term interest rate
which is said to influence investment and
aggregate demand (however do businesses
and households now borrow on a variable
interest rate?)
• Monetary policy is a blunt instrument, with
long and variable lags.
Some additional critiques: Efficacy of
Monetary Policy
•
•
Real interest rate hikes may lead to higher inflation
rates, through interest cost push; it also generates
an additional flow of interest payments in a stockflow consistent model
Empirically, evidence suggests that the interest
elasticity of investment is non-linear and
asymmetric.
Interest rates ↑ → investment ↓ in times of
economic booms, the reverse is not true.
Interest rates ↓ → unlikely investment ↑ in times of
recession.
What about the role of fiscal policy?
• Is counter-cyclical fiscal policy dead?
• What about the US with Reagan and Bush?
• What about Europe and its fiscal constraints
imposed by the Maastricht Treaty (deficit at a max
of 3% of GDP)?
• Is Canada a counter-example of Keynesian
economics (growth picked up when the federal
government cut into its expenditures and raised
taxes)? Is there an explanation of this
phenomenon? (there is!).
The Kriesler/Lavoie Phillips Curve
Inflation rate
ufc : full capacity
utilization
πn : rate of inflation
associated with the
normal range of output
πn
um
Ufc
Rate of capacity
utilization
Post-Keynesian Phillips Curve
In the long run ….
Inflation rate
Utilization rate
um
ufc
Empirical work that gives support to a
horizontal Phillips curve
• Eisner 1995, 1996, identifies a traditional segment,
followed by a flat segment, even with high rates of
utilization
• Filardo, Kansas City Federal Reserve Bank Economic
Review, 1998, has the three segments of PUP.
• Driscoll and Holden (2004).
• Most recent studies at the Bank of Canada cannot find a
relationship between the rate of utilization (or the output
gap) and the inflation rate: the Phillips curve looks
everywhere flat.
• Bloch and al. (JPKE, 2004) pretend that strong world
demand for raw materials is the main source of inflation,
and not domestic demand within each country.
Conclusion:
• Even if natural growth rates are not endogenous, it
is very likely that current monetary policies have
led to overly low rates of utilization and overly
high unemployment rates.
• Central banks tell us that they are doing a good
job. But lower interest rates could have been set,
without inflation rising. In other words, high rates
of capacity utilization may not have the
inflationary effects that central banks presume
they have.
…And this is recognized by some
(previous) central bank officers
• « If expectations of low inflation are firm, and the Phillips curve is
almost flat over a wide range at 2 percent inflation, it may be time to
make monetary policy explicitly accountable for the cyclical
component of output. To implement the idea, following the clause
• “Within this range monetary policy will continue to aim at keeping the
trend of inflation at the 2 per cent target midpoint”
• the official statement could add
• “Monetary policy will seek the maximum levels of output and
employment consistent with this target.” »
• K. Clinton (Bank of Canada 1972-2003), Bordeaux 2006.
A first addendum: financial crises
• Whereas in normal times, all short-run interest
rates move in tandem, this is not always the case,
especially in times of financial crises, as we have
experienced since August 2007, as a result of the
subprime crisis.
• Over a certain number of months, rates on
commercial paper have exceeded the overnight
rate, while rates on Treasury bills are way below
the overnight rate
A Minskyan new consensus
• Thus we must add a quadrant to our little
model, carefully distinguishing the
overnight rate (the fed funds rate f) from the
rate which is relevant to the private sector,
which is now called r.
• Thus the reaction function determines f, but
r may not move in tandem with f.
Adding another rate of interest to the
four quadrant story
• Assume that the economy starts off from point A, shown in all
quadrants of the Figure. In a ‘Minsky moment’, the risk spread τ rises
considerably. As there is a rush towards liquidity and riskless assets,
the prices of risky assets fall, and hence the interest rates on these
assets rise. This is represented in Figure 7 by an upward shift of the
transmission mechanism curve, from TM1 to TM2. Thus, even if the
financial crisis does not have a direct impact on the real economy –
meaning here that the parameters in the IS equation remains put, there
will be consequences for the real economy because market rates will
rise from r1 to r2. If the central bank does not modify its reaction
function, the aggregate demand curve will get shifted downwards from
AD1 to AD2 and economic activity will fall to u2, with the economy
moving from point A to point B. Then, as inflation rates start to fall, as
a result of the rates of capacity utilization being below their normal
levels, nominal and real federal funds rate will be brought down by the
central bank, moving along the RF1 reaction function.
Figure 7
TM2
r
A’

TM1


r
r2
B

A

r1
A
IS
f
B
A’ A
B
RF1
BT
RF2


f1

A
B
B2
f
f2
u
C
u2

un
IA
AD2 AD1
u
Other rates do not follow the overnight rate
(the target is 2%)
Treasury bill
1-month rate
• 09/12/2008:
• 09/15/2008:
(Lehman Monday)
• 09/17/2008:
• 09/18/2008:
• 09/19/2008:
• 09/22/2008:
• 09/23/2008:
Corporate non-financial
1-month rate
1.64%
0.36%
0.07%
0.26%
0.75%
0.76%
0.30%
• 09/12/2008:
• 09/15/2008:
(Lehman Monday)
• 09/17/2008:
• 09/18/2008:
• 09/19/2008:
• 09/22/2008:
• 09/23/2008:
2.39%
2.47%
2.51%
2.55%
2.63%
3.02%
3.38%
Higher risk premia act as a real
negative shock on the economy
• If the monetary authorities react in the normal way,
without taking into account the turmoil in the financial
markets and the higher risk premia, the economy will be
brought back to the normal rate of capacity utilization, but
at a steady rate of inflation that will be below the target, at
point C. In other words, the risk premium in financial
markets acts like a negative shock on the aggregate
demand curve, just like a real negative shock would. If this
financial shock is large enough it could even bring about
negative inflation, and hence debt-deflation as argued by
Minsky.
•
The central bank must counter the
risk premium shock
• The higher risk premium requires the central bank to modify its
assessment of the natural rate of interest, f0, in a counter-intuitive way,
since it must reduce the federal funds rate when long-term market
interest rates are rising. The monetary authorities must shift their
reaction function, to RF2, and reduce the federal funds rate to f2 in
order to keep inflation on target at πT and avoid a reduction in
economic activity (thus keeping market interest rates at r1). If they do
so, and assuming the financial turmoil does not have further effects,
the central bank could succeed in keeping the economy at the normal
rate of utilization, thus keeping the economy at points A in the northeast and south-east quadrants, which correspond to points A’ of the
north-west and south-west quadrants.
A critique of the way some central bankers
claim to assess the natural rate of interest
• The central bank cannot claim, as it
sometimes does, that long-term market rates
are a proper proxy for the natural interest
rate and a guide to help setting the federal
funds rate. In the present case, if the central
bank were to raise its estimate of the natural
rate of interest, in an effort to follow the
apparent increase in long-term market rates,
it would only make matters worse.