Transcript 7.Monetarism_to_Cambridge.pps

Lecture Seven
The Breakdown of Keynesianism
The Rise and Fall of the Phillips Curve
The Neoclassical Revival
The Keynesian Counter-attack:
logical flaws in Neoclassicism
Recap
• IS-LM dominates interpretation of Keynes
• Keynes a “marginalist”: liquidity preference as marginal
cost of interest foregone in holding money
• Role of expectations under true uncertainty ignored
• Neoclassical synthesis: combination of IS-LM macro with
Walrasian micro
Breakdown of Keynesianism: “Philips Curve”
• 1958 study identifies relationship between unemployment
& rate of change of money wages
• Consonant with Keynes on real wages and employment
– Inc. output --> decrease real wages --> prices must rise
• “Inflation-unemployment trade-off” perceived as core of
Keynesian policy
• 1960s--accelerating inflation. “Breakdown” of Philips’
Curve opens door to Friedman’s monetarism
The Phillips Curve
• 3 factors which might influence rate of change of money
wages:
– Level of unemployment (highly nonlinear relationship)
– Rate of change of unemployment
– Rate of change of retail prices “operating through cost
of living adjustments in wage rates… when retail prices
are forced up by a very rapid rise in import prices … or
… agricultural products.” [Economica 1958 p. 283-4]
• Cost-based perspective on prices
• Developed curve from UK wage change/unemployment
statistics from 1861-1913
• Only 1st of 3 causal factors shown in curve
The Phillips Curve
• Found a “clear tendency” for
– inverse relation between U and rate of change of
money wages (Dwm)
 Dwm above curve when U falling, and v.v
• Fitted exponential curve to data:
y  a  b. x
Dwm
c
Unemployment
log y  a   log b  c.log( x )
log y  0.9 .984  1394
. .log( x )
The Phillips Curve
Deviations from trend because of:
Fitted through average wage
change & U for 0-2,2-3,3-4,
4-5,5-7,7-11% unemployment
Wage-price spiral
due to wars; falling U
Rising unemployment
The Phillips Curve fitted to 1913-1948 data
War-induced
rise in M
prices
Rapid rise in U;
13% fall in M prices;
“cost of living” agreements
The Phillips Curve: 49-57 data with time lag
Close fit
of 50s UK
data to
curve
Import price rise
The Phillips Curve
• Conclusion: Phillips extrapolates from money wages to
price inflation:
– “Ignoring years in which import prices … initiate a
wage-price spiral, which seems to occur very rarely
except as a result of war, and assuming an increase in
productivity of 2% p.a., … [for] a stable level of
product prices … unemployment would be … 2.5%. [For]
stable wage rates … about 5.5%” [p. 299]
– An inflation-unemployment trade-off?
• But Phillips’ main purpose for developing it was to
provide an input for his dynamic models in which
unemployment, output, etc., varied cyclically.
• Nonetheless, trade-off interpretation becomes part of
orthodox Keynesianism
The Phillips Curve: Breakdown…?
OPEC I
Vietnam
war
OPEC II
The rise of monetarism
• Friedman’s explanation: adaptive expectations, the short
run trade-off but long run vertical Phillips curve
• Basic model:
– Exogenously given money-supply (“helicopters”)
– Walrasian economy in long-run equilibrium: all prices
currently market-clearing
– No sale of capital assets possible
– Static-stochastic economy: no growth, aggregates
constant, but random disturbances to individuals
• “Aggregates are constant, but individuals are subject to
uncertainty and change. Even the aggregates may change
in a stochastic way, provided the mean values do not.”
[OREF II 119]
Monetarism
• Basic model (cont.)
Back to Hicks’s pre-Keynesian
– Motives for holding money: “typical classical theory”
• Transactions (barter motive)
• “uncertainty” says Milton, but: means of aggregates
constant, stochastic variation only? Risk, not
uncertainty.
– No variation in parameters of risk considered
– Holdings of money related to level of transactions:
• Md = k.I
– Exogenous increase in Ms-->
• initial increase in output but constrained by already
fully-employed economy
• price level bid up till “real value” of money holdings
restored
Monetarism
• Continuous growth in Ms?
– (corresponds to policy for lower U under Phillips curve
inflation/U tradeoff)
– “which, perhaps after a lag, becomes fully anticipated
by everyone”; adaptive expectations [OREF II]
– results in…
• Adaptive Expectations (with certainty)
– “what raises the price level, if at all points markets
are cleared and real magnitudes are stable? ... Because
everyone confidently anticipates that prices will rise…”
[OREF II]
– Increasing Ms raises prices, no impact on output in long
run (short run impact until expectations adapt)
Adaptive Expectations and the Phillips Curve
“Long
Target
run
Rate
Phillips
Curve”
Accelerating
Short run gain
inflation needed with long run
to sustain target
pain...
DMs causes some growth but…
Expectations adapt
Expected Inflation=
DMs- DLab.Prod
Expectations adapt
Economy returns to pre-existing “natural”
rate
SRPC3
SRPC2
Unemployment
Initial “natural” U rate with zero expected inflation
SRPC1
From Monetarism to Rational Expectations
• Friedman adaptive expectations
– expectations adapt to change after experience of
inflation caused by increased money supply
– short-term impact of policy neutralised in long term
• Rational Expectations (Muth/Sargent): expectations
predict consequence of change based on rational model of
reality:
– “expectations … are essentially the same as the
predictions of the relevant economic theory.” [OREF
III]
• Combined with lag formulae as explanation for cycles:

See chaos
theory,
later
p e t  V j pt  j
j 1
Expected price in cyclical market
a weighted sum of previous prices
Rational Expectations Macro
• “the public knows the monetary authority’s feedback rule
and takes this into account in forming its expectations…
unanticipated movements in the money supply cause
movements in y [output], but anticipated movements do
not.” [OREF III]
• “Predictions of relevant theory” are: increased Ms will
increase price level --> instant adjustment of prices to
government Ms policy --> no impact on output
• “Natural rate” of unemployment + rational expectations =
policy ineffectiveness hypothesis:
Rational Expectations Macro
• “by virtue of the assumption that expectations are
rational, there is no feedback rule that the authority can
employ and expect to be able systematically to fool the
public. This means that the authority cannot expect to
exploit the Phillips curve even for one period.” [OREF]
• A vertical Phillips curve in short run
• Prediction that policy could never have been effective
• Short-run movements in unemployment due to
unanticipated shocks
Problems for Rational Expectations
• (1) Was Keynesian policy “ineffective”? (see next table)
– By RE, differences between Keynesian & Neoclassical
policy periods should be
• Higher M growth, Inflation
• No difference in unemployment
• (2) Why the Great Depression?
– “Natural rate” moves by 30%???
• (3) Movement of “natural rate” (hysteresis?)
• (4) Nature of expectations
– Keynes: behaviour under fundamental uncertainty
– RE: behaviour under certainty, using economic theory
to predict
The Keynesian/Neoclassical Scorecard
USA Economic Policy Scorecard
Keynesian Neoclassical
(50-73)
(74-96)
M1 Change
4.36%
6.79%
Int Rate
3.39%
7.06%
GDP Inc.
3.90%
2.34%
U Rate
4.81%
6.92%
CPI
2.45%
5.81%
Outcomes contradict RE Hypothesis: Keynesian period has Lower
Money Growth, Inflation, Interest Rates, Unemployment Higher
Growth
How could this be if policy was “ineffective”?
Did the economy suddenly deteriorate? (possibly true…)
Theoretical Developments
• Behind policy fight, theoretical battles
– Perfection of Walrasian GE model (Arrow-Debreu)
– Classically-inspired critique of neoclassicism (Sraffa)
• Arrow-Debreu General Equilibrium
– proof of existence/uniqueness/stability/optimality of
equilibrium (completing Walras’ work), but
– under assumptions of given resources, given tastes,
given possible future states of world, “contingent
contracts”, and no uncertainty, and peculiar definition
of commodities
• The Keynesian Critique: reality that “factors of
production” are commodities, labour, land applied to show
that neoclassical theory is internally inconsistent
The “Cambridge Controversies”
• Critique of neoclassical economics initiated by Joan
Robinson, Piero Sraffa (Cambridge UK) over nature of
capital vis a vis land, labour
• Theory defended by Paul Samuelson, Robert Solow
(Cambridge USA)
• Focus of attack: validity of neoclassical theory of
distribution based on supply and demand
The “Cambridge Controversies”
• Neoclassical theory argues that
– increasing supply of factor of production will
reduce its price
– reducing its price will increase its use in production
– Factor’s price equals its marginal product
– Direct relationship between supply of factor and
its price
– Models production as
• involving “factors of production” (Land, Labour,
Capital) as inputs and goods as outputs
– versus classical position: goods produced using
goods and labour as inputs
• The neoclassical position of profit and capital is…
The “Cambridge Controversies”
Labour
Output
Increasing
supply…
Decreasing price...
Marginal Product
Capital
Diminishing
marginal
product
Capital
Capital
Increasing use of factor
relative to others...
Rate of profit is the
marginal product of
capital…
The “Cambridge Controversies”
• Sraffa, 1960
– Take economy in full general equilibrium
• All “marginal” changes complete
– What determines prices in full equilibrium if all
marginal changes are over?
– Self-reproducing system of commodity production
• inputs commodities & labour
• output commodities
• equilibrium prices of outputs must enable their purchase
as inputs in next period
– System (1): Simple reproduction, commodity inputs
only:
The “Cambridge Controversies”
• 240 qr wheat + 12 t iron + 18 pigs --> 450 qr wheat
• 90 qr wheat + 6 t iron + 12 pigs --> 21 t iron
• 120 qr wheat + 3 t iron + 30 pigs --> 60 pigs
• 450 qr wheat | 21 t iron | 60 pigs (sum of inputs=sum of
outputs)
• Regardless of demand, prices must allow system to
reproduce itself:
– 450 qr wheat must buy 240 qr wheat, 12 t iron, 18 pigs
– 21 t iron must buy 90 qr wheat, 6 t iron, 12 pigs
– 60 pigs must buy 120 qr wheat, 3 t iron, 30 pigs
The “Cambridge Controversies”
240  pw  12  pi  18  p p  450  pw
System of production:
90  pw  6  pi  12  p p  21  pi
120  pw  3  pi  30  p p  60  p p
As a matrix
 240

 450
equation:  90
 21
 120

 60
Has the solution:
12
450
6
21
3
60
18 

450   pw   pw 
12     
 p    p 
21   i   i 
30   p p   p p 

60 
 1
 pw   
   10 
 pi    1   $
   1
 pp 
 2
i.e., price system for simple reproduction independent of
demand, marginal utility, etc.; depends instead on system of
production
The “Cambridge Controversies”
• System (2): Expanded reproduction
• surplus produced, must be split between capitalists
and workers
– in equilibrium, a uniform rate of profit r, uniform
wage w
A
A
a
b
 pa  Ba  pb ... Ka  pk   1  r   La  w  A  pa
 pa  Bb  pb ... Kb  pk   1  r   Lb  w  B  pb
Amount of ...
Amount of B produced
A used to
Ak  pa  Bk  pb ... Kk  pk   1  r   Lk  w  K  pk

produce B
Labor fully employed
La  Lb ... Lk  1
Aa  Ab ... Ak  A, Ba  Bb ... Bk  B...
+ive net output
The “Cambridge Controversies”
• r & w values determine split of surplus between
capitalists, workers. To determine prices, must therefore
know either r or w beforehand
– Distribution therefore not determined by “market”
– Instead, different pattern of prices for every pattern
of distribution: marginal productivity theory of income
distribution incorrect in general equilibrium
• But what about validity of production function, isoquants,
when marginal changes still relevant?
The “Cambridge Controversies”
• Neoclassical position (by Samuelson):
– Concedes Classical position more factual
• output produced by heterogeneous commodities and
labour, aggregate capital an abstraction
– But neoclassical position still defensible as an
abstraction
• Samuelson (for neoclassicals) argues
– isoquants just a parable we use to teach students
– reality is different technologies, each with fixed ratio
of capital to labour
– increase in price of capital will lead to less capital
intensive technology being chosen:
Labour
The “Cambridge Controversies”
Technology 1: low K/L
ratio, used when K
expensive
“Envelope” is isoquant
Technology 5: high K/L
ratio, used when K cheap
Capital
Decreasing price of capital means more
capital intensive methods used, akin to
simple parable that decreased price means
more capital used
The “Cambridge Controversies”
• As an aside, Samuelson ridicules classical theory’s
problems with labour theory of value, that capital-labour
ratios must be the same in all industries.
• Problem: Samuelson assumes each technology can be
represented by a straight line relationship between
capital and labour
• Garegnani shows that straight line relationship only
applies if capital to labour ratio is the same in all
industries
• If K/L ratios differ, each technology will be represented
by a curve, not a straight line
• Curves can cut each other in more than one place:
Labour
The “Cambridge Controversies”
Technology 1: low K/L
ratio, used when K
expensive
“Envelope” is isoquant
Technology 2: only used in
intermediate K/L price
range
Technology 1: could also be
used when K cheap
Capital
Problem known as “reswitching”: simple neoclassical parable
does not work when multiple industries considered.
The “Cambridge Controversies”
• Why a curved relationship?
– The definition of capital
• What is “capital”?
– Money?
– Machine?
• Both, obviously; but how to “add” machines together?
• Money value only common feature
– but money value reflects expected profit
– “rate of return” and “value of capital” thus linked
• Sraffa’s solution: reduce all machines to “dated labour”
– Machine today produced with
• labour last year, plus
• machinery inputs last year
The “Cambridge Controversies”
• If economy has been in long run equilibrium for indefinite
past
– then all goods produced earned normal rate of profit r
– therefore value of machine now equals
• value of previous year’s inputs (labour and capital)
• multiplied by 1+r
– Do it again: replace last year’s machine inputs with
• labour and capital used to produce those machines
• multiplied by 1+r
– Get a whole series of terms for the labor input each
year multiplied by 1+r, (1+r)2, (1+r)3
– Machine/commodity component reducible to almost
(but not quite) zero.
The “Cambridge Controversies”
• Next: the “standard commodity”
– Earlier, Sraffa shows how to devise a “measure of
value” unaffected by the distribution of income: the
“standard commodity”
– When measured using this, there is a simple linear
relationship between the real wage w, the rate of
profit r, and the maximum possible rate of profit R:
r  R  1  w
This can be reworked to give an expression for the wage
in terms of the rate of profit:
r

w  1  
R

The “Cambridge Controversies”
• Each year’s labor input to producing a machine today is
thus broken down into
– the number of units of labor performed (say 1 unit)
– times the wage rate w (now expressed in terms of r &
R)
– times 1+r raised to the power of n, for how many years
ago the labor was applied:
Number of years
Wage in
r
ago that machine
n

terms of rate
1    1  r
was made
R


of profit
Rate of profit
Expression gives an unambiguous value for today’s capital
input in terms of dated labor, but…
the measured value of capital depends on the rate
of profit:


The “Cambridge Controversies”
• So rather than the rate of profit depending on the amount
of capital (marginal product theory of income distribution),
the amount of capital depends on the rate of profit
• Second problem: this relationship is very nonlinear
– First part falls uniformly as rate of profit rises
– Second part
• rises slowly as r rises
• rises rapidly as n (number of years ago rises)
– Two parts interact very unevenly
• For small change in r, second effect outweighs first as n
rises
• For large change in r, first effect outweighs second for
small n
• In between, can’t pick whether increasing r will increase
or decrease measured amount of capital:
The “Cambridge Controversies”
Value of machine produced with one unit of labor applied n
years ago at a rate of profit r between zero and 25% when
R=25%:
Made this
year (n=0)
1
w( r , 0) 0.5
1
Made 5
years ago
(n=5)
w( r , 1) 0.5
r=0
r=25%
0
0
0.05
0.1
0.15
0.2
0
0.25
0
0.05
0.1
r
w( r , 10)
20
1.5
15
Made 10
years ago
(n=10)
0.5
0
0
0.05
0.1
0.15
r
0.2
r
2
1
0.15
0.2
w( r , 25)
Made 25
years ago
(n=25)
10
5
0.25
0
0
0.05
0.1
0.15
r
0.2
0.25
Measured
value rises
& then
falls as
rate of
profit
rises
0.25
The “Cambridge Controversies”
• Can’t apply marginal productivity theory to capital:
– return to capital can’t reflect marginal product of
capital
• measured amount of capital depends on rate of profit
– numerator/y-axis (r) and denominator/x-axis
(“amount of capital”) are interdependent
– relationship is “messy”
• rises as r rises for a while
• then falls as r rises
Output
– Rate of profit therefore can’t be “marginal
productivity of capital”
Diminishing
marginal
product
Capital
The “Cambridge Controversies”
• Numerous other facets to Cambridge Controversies
• Minority of neoclassicals who got involved in debate
(Samuelson, Solow, Hahn, etc.) had 2 eventual responses
– Grudgingly conceded critique had validity and started
to develop alternative approaches to neoclassicism
themselves (Samuelson, Solow)
– Abandoned attempt to make neoclassical economics
relevant to real world and developed general
equilibrium models as abstract thought experiments
only (Hahn, etc.)
• Majority of neoclassicals assumed (wrongly) that debate
won by neoclassicals and continued on as always.
• Raises issues of methodology (is economics a science?)
discussed in 2 weeks. Next week, finance...