After Keynesian Macroeconomics

1
After Keynesian Macroeconomics

• An article by Robert E. Lucas and Thomas Sargent

2
The Keynesian Failure of the 1970s

• In 1960s, expansionary fiscal policies led to
  prosperity during Kennedy-Johnson era, suggesting
  validity of Keynesian theory.

• However, in 1970s, economy experienced heavy
  stagflation, clearly contradicting Keynesian
  theory and calling for its reevaluation.

3
Why did the Keynesian Model Fail?

• Structural models have isolated parameters that
  remain constant across various shocks, such as
  changes in economic policy.
• Keynesian models not structural use historical
  data to estimate model parameters and then use
  parameters to predict policy consequences.
• Empirical tests verify that Keynesian models only
  forecast accurately over short-term,
  unconditional periods.

4

• Keynesian Models also have 3 major
  oversimplifying a priori restrictions
• (1) Assume many explanatory variable coefficients
  equal 0. (i.e. model for money demand depends
  only on output and interest rates, excluding
  variables that may matter like future output).
• (2) Assume pure error terms that have no
  heteroskedasticity, serial correlation, or
  correlation with dependent variable.
• (3) Assume exogeneity for certain variables, like
  government spending, which may be determined
  within model.

5
An Example of Oversimplification Expectations

• Supply and demand functions depend on peoples
  expectations of income, price levels, and tax
  rates.
• Keynesian models assume that an expectation
  variable depends only on past values of the
  variable itself. But, if we believe that people
  act

intelligently and in their own self interest,
people must base their expectations on a variety
of variables.

• Though including these expectations relationships
  would complicate the model, there is no
  theoretical basis for leaving them out.

6
Consequences of the Keynesian Failure

• Increased openness less refining of Keynesian
  models and more exploration of alternative
  business cycle models.
• Distrust quantitative/scientific methodology
  prefer judgmental methodology.
• In response, L S suggest equilibrium business
  cycle theory.

7

• Keynes founded macroeconomics to explain
  characteristics of business cycle
• Length and severity of depressions as well as
  extensive unemployment could not be reconciled
  with the two microeconomic postulates
• markets assumed to clear
• agents assumed to act in self-interest

• Classical model could not explain positive
  correlation between prices/wages and aggregate
  output/employment
• Keynesian models imply causal connections from
  aggregate demand and inflation to output and
  employment contrary to classical dichotomy

8

• Relaxing assumption of perfect information can
  explain correlations while maintaining classical
  assumptions
• Agents make best possible estimate of all
  relative prices affecting supply/demand decisions
• Agents may sometimes mistake unexpected inflation
  for an increase in the relative price of a good
  theyre selling, causing them to overproduce
• Rational expectations assumed agents make best
  possible use of limited information
• Theory accounts for positive correlation between
  unexpected changes in price level and aggregate
  output

9

• Correlations DO NOT depict tradeoffs
• Theory predicts that there is no way that the
  monetary authority can implement a successful
  countercyclical policy
• Impossible for government to make systematic
  changes that are not anticipated to the money
  supply
• Model explains correlation while adhering to
  classical postulates, but implies non-activist
  monetary policy

10

• Rational expectations
• Cross-equation restrictions among parameters
• If MS affects ?e and ?e affects I, then MS
  affects I

• Variable forecasts depend on many variables, not
  just lagged values of the variable in question
• If variable y helps predict variable x, implies
  that x cannot be regarded as exogenous with
  respect to y

11

• The Lucas Critique
• Policy intervention flows through to all
  equations in model because of cross-equation
  restrictions
• Change in policy affects behavior changes by
  public
• Estimated coefficients are no longer constant
  after policy change
• Football example
• If granted five downs, teams wouldnt punt on 4th
  down

12

• Criticism Cleared Markets
• Uncleared labor market seemingly contradicts
  theory of cleared markets
• Long-term labor contracts with two-three year
  horizons exist (Fischer model)
• Long-term contracts dont necessarily imply
  rigidity or invalidate cleared markets
• Can adjust according to cost of living, etc.
• Affected by monetary policy (Lucas Critique)

13

• Criticism Persistence
• Model suggests that agents forecast errors are
  random and serially uncorrelated and thus cannot
  account for long business cycles
• Several propagation mechanisms create serially
  correlated booms and busts

Firms adjust capital and labor gradually to
minimize costs of responding to relative price
signals Households divide changes in income
between immediate and future consumption. Search
theory workers dont necessarily take first job
offer, firms dont hire first applicant
14

• Criticisms Linearity and Stationary Models
• Equilibrium models assume linearity without any
  theoretical reasoning
• Non-linear models likely wouldnt change results
• Models assume agents have been operating for a
  long time in a stochastically stationary
  environment
• Models do not account for learning
• Agents are instantly smart and know the
  probability distributions they face
• Cumbersome, and wouldnt distinguish equilibrium
  models from Keynesian

15
Benjamin Friedmans Response

• L S argue that Keynesians assume rules of
  thumb whereas classicals (equilibrium business
  cycle theory) assume optimizing behavior.
• Friedman rejects this idea, providing examples
  of Keynesian theories based in optimizing
  behavior, such as the life-cycle model of
  consumer behavior.

• Also argues that equilibrium business cycle
  theory is no less guilty of making arbitrary
  restrictions than Keynesian theory. Uses the
  Islands theory to make this point.

16
The Islands Example

• If a shoemaker in an economy with rising general
  prices sees that the prices of shoes are rising,
  he might mistakenly increase production of shoes,
  assuming he has no knowledge of relative prices.
• However, what if the shoemaker sees the price of
  leather rising before he sees the price of shoes
  rising? In that case, he would mistake this for
  a relative increase in an input and reduce
  production.

• The assumption that producers see output prices
  before input prices is, therefore, arbitrary.

17
Friedmans Other Concerns
Equilibrium business cycle theory describes the
long-run it cannot be applied to questions
people have about economic policy in the
short-run because expectations do not adjust
instantaneously.

• Lucas and Sargent ignore that consistent fiscal
  policy in the U.S. decreased unemployment from 9
  percent to 6 percent between 1975 and 1978 while
  inflation increased. In Europe, where fiscal
  policies were extremely different, unemployment
  remained high and inflation remained low.

18
Reply and Rebuttal
L S response dismisses Friedmans points as too
general and mainly defends own rhetoric If one
is to discuss this well-documented discrepancy
failure of 1970s what language is appropriate?
Should these forecasts be termed accurate or an
econometric success?

• Friedmans Rebuttal continues to critique L S
  rhetoric and their lack of specificity Why not
  say precisely which models are under criticism
  and then look carefully at their actual record of
  performance? That, and certainly not any resort
  of wishy-washy verbal compromise, is the
  alternative I suggest in place of the unfocused
  rhetorical attack on Keynesian models that
  Professors Lucas and Sargent presented here.

19
Discussion Questions

• Our intent will be to establish that the
  difficulties with the Keynesian failure of the
  1970s are fatal that modern macroeconomic
  models are of no value in guiding policy, and
  that this condition will not be remedied by
  modifications along any line which is currently
  being pursued. How damaging was stagflation in
  1970s to Keynesian model? Is the Keynesian model
  fatally flawed?
• Do you believe that the ineffectiveness of
  monetary policy can be applied in the short-run
  or is it only a long-run phenomenon? Do people
  adjust their expectations quickly enough in the
  short-run?
• Do you believe that equilibrium business cycle
  theory is more grounded in optimizing assumptions
  than Keynesian theory? Do you agree assumption
  of Island theory is arbitrary?
• L S argue that policy changes shift model
  coefficients. Is this still a problem? Is it
  possible that we have had enough different types
  of monetary policies since the 1970s that we have
  enough historical data to adjust our model
  coefficients to reflect appropriate changes in
  policy?