Lecture V New Classical Economics

Download Report

Transcript Lecture V New Classical Economics

Lecture V New Classical Economics and Robert Lucas
Attack on Keynesianism (again)
• Like the monetarists before them, the new classical
economists also set their sights on Keynesian theory,
which they view as simplistic and damaging to the
economy
• Quite opposite to the monetarists, however, new classical
theory restates the classical dichotomy of the neutrality
of money (although unanticipated monetary shocks can
have real affects)
• The rational expectations theory used by new classical
economists replaces adaptive expectations, which will
lead to consistently biased estimates of economic
variables
• The economist most closely associated with new
classical theory is Robert Lucas, Jr.
Rational expectations
• The basis of the new classical approach is the idea that
expectations are formed “rationally”
• It has been argued that it was a stroke of genius to declare your
theory to be “rational”; after all, by implication any alternative
theory must be irrational, and no one wants to go there
• By rational expectations we mean that on average economic
agents correctly perceive economic variables
• Thus, for example, on average workers will accurately guess the
inflation rate next period and will act in such a way as to protect
themselves from any adverse effects this will cause them
• It does not mean, as some have claimed, that predictions are
always correct or that everyone’s predictions are accurate
Origin of rational expectations
• The seminal paper on rational expectations was John
Muth’s 1961 Econometrica article Rational Expectations
and the Theory of Price Movements
• Lucas used the theory of rational expectations to account
for the apparent non-neutrality of money
• According to his formulations, in a world of money
surprises, real effects can occur but only to the extent
that these surprises are unanticipated
• He then showed operationally how to decompose
monetary shocks into two components: an anticipated, or
forecastable component, and an unanticipated
component
• Only the unanticipated component should have real
effects
Other adopters of RET
• Thomas Sargent and Neil Wallace used RET to
develop the Policy Ineffectiveness Proposition,
or PIP, which shows that if agents use all
information efficiently and expectations are
realized, on average, then discretionary policies
will quickly lose their ability to affect real
economic variables
• Even New Keynesians, such as Stanley Fisher,
adopted RET in their economic models
The simple RET model
• Let us suppose we wish to forecast the rate of inflation p
• We start with the assumption that economic agents use any
and all information available to forecast next period’s inflation
rate, which, it is assumed, has one unique value
• If the forecast is unbiased then p = p* + u and E(p) = p*,
where p is the forecast value of inflation, p* is the actual
value, and u is random error with E(u) = 0 and is independent
of p*
• An alternative, more realistic model, assumes that each
agent uses information up to the point that the marginal cost
of the additional information equals the marginal benefit; this
is the weak form of the so-called efficient market hypothesis
• The assumption that all information is used is the strong form
of EMH
Criticism 1
• Model assumes a unique p*; in the real world multiple p*
may be possible and which p* occurs is based on
agents’ actions
• For example, the Fed announces a 5% target inflation
rate, which is now 10%
• if people do not believe the Fed and expect 10%
inflation, wage rise 10% and inflation is 10%
• If the Fed, however, cuts money growth to coincide with
their target 5%, a recession follows
• On the other hand, if the Fed credible, then the expected
p* is 5% and wages and prices rise by 5% and there is
nor recession
• Thus, economic agents’ actions may determine p*
Criticism 2
• It is well known that assumptions about individual
behavior do not carry over to aggregate behavior
• Thus, even if all households and firms are rational
(and don’t believe the Fed will hit its 5% inflation
target in the previous example) they may still not
act on those beliefs
• So households, in order to be sure they don’t lose
their jobs IF the Fed meets the target, they may
accept a 5% wage increase
• And firms may only raise their prices 5%, just in
case the target is met so as not to have their prices
undercut by other firms
Credibility
• As noted in both of the previous examples,
people’s perceptions of whether or not the
central bank will be abel to “stick to its guns” and
hold the line on inflation, is critical in whether or
not people act in a way to achieve the desired
equilibrium
• Much of the literature today concerns itself with
the issue of central bank credibility, which is
based on central bank independence, or CBI
Mark I Example
• In Lucas’ early work, his models were based on
unanticipated nominal shocks as being the
primary source of macroeconomic instability; the
was referred to as monetary equilibrium
business cycle theory, or MEBCT
• As noted earlier, only the unanticipated
component of nominal changes could impact
real variables, as agents will shield themselves
against the anticipated component
A Lucasian Model
• Yy = YNt + YCt, where YN is the secular component of GDP
and YC is the cyclical component
• YNt = λ + φt is the underlying trend growth path of the
economy
• The cyclical component depends of the surprise and last
periods deviation from its natural rate
• YCt = α[Pt – E(Pt|Ωt-1)] + β(Yt – YNt-1)
• Ω represents all variables used to predict P, and β > 0
determines the speed with which output returns to its natural
rate after a shock.
• To explain why output and employment remain persistently
above or below their trend values for a succession of time
periods, further assumptions are made about the propagation
mechanism.
• These make reference to lagged output, investment accelerator
effects, information lags and the durability of capital goods,
contracts that inhibit immediate adjustment and adjustment
costs. In the filed of employment, firms may face costs both in
hiring and firing labor, interviewing and training new
employees, making redundancy payments, etc. [Later we shall
see that New Keynesians refer to many of the same factors in
their models]
• As a result firms may adjust employment and output gradually
over a period of time after an unanticipated shock
• Combining we get Yt = λ + φt + α[Pt – E(Pt|Ωt-1)] + β(Yt – YNt1) + εt, where εt is a random error process
• To account for how accustomed agents in a country are to
demand shocks Lucas adds a factor θ to account for speed of
adjustment as follows:
• Yt = λ + φt + θα[Pt – E(Pt|Ωt-1)] + β(Yt – YNt-1) + εt, the larger
the θ, in a country that has not had many inflation shocks, the
greater the output impact of price shocks. For example, in
countries with lots and variable rates of inflation COLAs will
be built into their contracts, so misperceptions are far less of
an issue than in a country that has not experienced much
inflation.
• In his model, the greater the variability in prices, that is, the
less that as attributed relative price variability, the smaller the
cyclical response of output to a monetary variation.
A PIP example; and increase
in the money supply
LRAS
SRAS'
P2
C
P1
SRAS
B
P0
A
AD'
AD
YN
Y
If the authorities announce and are believed
• Suppose the authorities announce they
will increase the money supply
• If agents believe the announcement, then
workers will insist on higher wages as AS
moves from AS to AS' and AD will shift
outward from AD to AD
• Aggregate output remains the same at YN
and prices rise directly from to P2.
Unannounced or not believed
• Now if agents do not believe the authorities, or if the
central bank does not announce the monetary increase,
then agents believe the increased wages and product
demand to be relative
• At first producers increase output, perhaps by paying
more for overtime hours, and consumers buy more
• This shifts AD outward to AD' and price rises to P1
• Then as workers and producers see the higher prices,
they insist on higher wages which can be paid at the
higher prices
Phillips curve
• In the first case, even in the short run, the
increased money supply had no effect on output
or prices
• In the second case, the results are the same as
under adaptive expectations; first an increase in
output without wage increases, then an increase
in wage rate and a decline in output to its
original level
• This is the story of the short-run versus long-run
Phillips curve
Mathematics of the reduced-form
Phelps-Friedman PC
p = inflation rate
U = unemployment rate
UN = natural rate of unemployment
S = exogenous supply shock
D = exogenous demand shock
m = money growth
pt = pet – φ(Ut – UNt) + φθSt
So Ut = UNt -1/ φ(pt - pet) +φθSt
pt = mt + θDt is the structural relationship between money and
prices, ie. MV=PY
• pet = met is rational expectation of the inflation rate
• mt = λ0 + λ1(Ut-1 – UN(t-1)) + θmt is the monetary
authority money growth response function; if
unemployment is above the natural rate, they will
increase money growth, andθmt is a random or
unanticipated money shock.
• met = λ0 + λ1(Ut-1 – UN(t-1)) is also the expected
inflation rate in period t
• so mt – met = θmt the only source of misperception
of the money growth rate for rational agents is the
monetary shock component; they correctly anticipate
the monetary rule of the monetary authority
Finally, PIP!
• The pt – pet = θDt + θmt So inflation misperceptions
are due to unanticipated demand and monetary shocks
• So Ut = UNt -1/ φ(pt - pet) + φθSt
= UNt -1/φ(θDt + θmt) + φθSt
• Thus, the systematic component of the monetary
growth is not included in the equation, so the
government’s policy rule has not influenced the rate
of unemployment
A Time Inconsistency game
• One of the arguments used by RET theorists is
that decisions may suffer from time
inconsistency
• Agents may find that their decisions are
governed as much, or more, by the timing of
them as by the wisdom of each alternative
• In this game, we imagine a central maker that is
fully aware of the consequences of
expansionary monetary policy; in the short run it
can reduce unemployment at the expense of
some inflation, but in the long run will leave
unemployment at its natural rate
Time inconsistency game graph
p
B
A
O
C
U
S2
S1
S0
S3
PCpe=0%
PCpe=2%
The game in action
• The banker starts with an economy at 6%
unemployment and no inflation; the Phillips
curve is PCpe=0% and lies on banker welfare
function S0
• If he inflates the economy along this short-run
Phillips curve, the optimal strategy is not to
announce the inflation and bring the economy
to point A on S1 > S0 where > implied higher
banker welfare
Rude facts
• Of course this is only temporary as agents come to
recognize that the inflation is general and not
relative; then they adjust prices and wage bringing
the economy from point A to point B with 6%
unemployment and 2% inflation
• Now, from this point, on S2 < S0 < S1, the optimal
strategy is to announce that he will disinflate back to
0% and for all agents to believe him and adjust their
wages and prices back to the original level
• In this case the economy returns to U = 6% and p =
0%
Final outcome of the game
• However, the banker has probably lost all credibility
as a result of his original inflating and instead of
believing him agents believe he will leave inflation at
its current level, hoping to return to point A
• Now if the banker carries through with his promise,
the economy follows the new SRPC from point B to
point C, where U = 7% and p = 0% along S1
• Thus, the optimal long-term strategy was to leave
inflation at 0% and accept a welfare of
So why play the game?
• Often, central bankers, because they lack either
the legal independence or the moral fiber, will
give in to pressure from the executive or
legislative branch of government and do the
wrong thing
• Again, central bank credibility is crucial in
avoiding the negative consequences of
disinflation; only if agents believe the monetary
authority is truly committed to inflation fighting
can it work
• Thus, CBI is essential