This Content Downloaded From 132.247.173.132 On Thu, 18 Apr 2024 18:20:37 +00:00
This Content Downloaded From 132.247.173.132 On Thu, 18 Apr 2024 18:20:37 +00:00
This Content Downloaded From 132.247.173.132 On Thu, 18 Apr 2024 18:20:37 +00:00
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
The University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend
access to Journal of Political Economy
Finn E. Kydland
Norwegian School of Economics and Business Administration
Edward C. Prescott
Carnegie-Mellon University
I. Introduction
473
best action, given the current situation, will not typically result in the
social objective function being maximized. Rather, by relying on some
policy rules, economic performance can be improved. In effect this is an
argument for rules rather than discretion, but, unlike Friedman's (1948)
argument, it does not depend upon ignorance of the timing and magnitude
of the effects of policy.
The reasons for this nonintuitive result are as follows: optimal control
theory is an appropriate planning device for situations in which current
outcomes and the movement of the system's state depend only upon
current and past policy decisions and upon the current state. But, we
argue, this is unlikely to be the case for dynamic economic systems. Cur-
rent decisions of economic agents depend in part upon their expectations
of future policy actions. Only if these expectations were invariant to the
future policy plan selected would optimal control theory be appropriate.
In situations in which the structure is well understood, agents will surely
surmise the way policy will be selected in the future. Changes in the
social objective function reflected in, say, a change of administration
do have an immediate effect upon agents' expectations of future policies
and affect their current decisions. This is inconsistent with the assump-
tions of optimal control theory. This is not to say that agents can fore-
cast future policies perfectly. All that is needed for our argument is that
agents have some knowledge of how policymakers' decisions will change
as a result of changing economic conditions. For example, agents may
expect tax rates to be lowered in recessions and increased in booms.
The paradox also arises in situations in which the underlying economic
structure is not well understood, which is surely now the case for aggre-
gate economic analyses. Standard practice is to estimate an econometric
model and then, at least informally, to use optimal-control-theory
techniques to determine policy. But as Lucas (1976) has argued, since
optimal decision rules vary systematically with changes in the structure
of series relevant to the decision maker, any change in policy will alter
the structure of these rules. Thus changes in policy induce changes in
structure, which in turn necessitate reestimation and future changes in
policy, and so on. We found for some not implausible structures that this
iterative procedure does not converge, and, instead, stabilization efforts
have the perverse effect of contributing to economic instability. For most
examples, however, it did converge, and the resulting policy was con-
sistent but suboptimal. It was consistent in the sense that at each point
in time the policy selected was best, given the current situation. In effect
the policymaker is failing to take into account the effect of his policy rule
upon the optimal decison rules of the economic agents.
In this paper, we first define consistent policy and explain for the
two-period problem why the consistent policy is suboptimal. The
implications of the analysis are then considered for patent policy and
The original objective of this research was to demonstrate the applicability of optimal
control methods in a rational-expectations world. We recognized the nonoptimality of the
consistent solution obtained by using control-theory techniques, but initially considered
this a minor problem. Further thought, in large part motivated by C. A. Sims's criticism
of our initial analyses, led us to the radical conclusions of this essay.
2 Uncertainty is not the central issue of this essay. As with Arrow-Debreu state-
preference theory, one need only define the decision elements to be functions contingent
upon observables to incorporate uncertainty as is done for the stabilization example in
Sec. V.
subject to
X1 = X1(it1, it2)
and
as ax2 as
S =2 + - = 0.
aX2 0a2 ait2
The consistent policy ignores the effects of 7i2 upon x1. For t
decision rule, the first-order condition is
as aX2+
as ax,
+
as+
+ a=
aaX2]
.
aX2 ait2 a
Only if either
effect of changes in x1 upon S both directly and indirectly through x2
is zero (i.e., [as/ax, + as/aX2 aX2/ax1] = 0) would the consistent policy
be optimal.
Pollak (1968) resolved a planning inconsistency which arose because
different generations had different preference orderings by assuming at
each stage that the policy selected was best (relative to that generation's
preferences), given the policies which will be followed in the future. For
the T-period problem, the 7IT is determined which, conditional upon
previous decisions git and xt, is best:
can be determined once future policy rules are known. With such a
procedure, the policy decision at each stage is optimal, given the rules
for future policy selection. 3 But as the simple example illustrated, this
procedure is suboptimal.
Two examples follow:
The issues are obvious in many well-known problems of public policy.
For example, suppose the socially desirable outcome is not to have houses
built in a particular flood plain but, given that they are there, to take
certain costly flood-control measures. If the government's policy were
not to build the dams and levees needed for flood protection and agents
knew this was the case, even if houses were built there, rational agents
would not live in the flood plains. But the rational agent knows that, if
he and others build houses there, the government will take the necessary
flood-control measures. Consequently, in the absence of a law prohibiting
the construction of houses in the flood plain, houses are built there, and
the army corps of engineers subsequently builds the dams and levees.
A second example is patent policy. Given that resources have been
allocated to inventive activity which resulted in a new product or process,
the efficient policy is not to permit patent protection. For this example,
few would seriously consider this optimal-control-theory solution as being
reasonable. Rather, the question would be posed in terms of the optimal
patent life (see, e.g., Nordhaus 1969), which takes into consideration both
the incentive for inventive activity provided by patent protection and the
loss in consumer surplus that results when someone realizes monopoly
rents. In other words, economic theory is used to predict the effects of
alternative policy rules, and one with good operating characteristics is
selected.
3There are some subtle game-theoretic issues which have not been addressed here.
Peleg and Yaari (1973) criticized Pollak's solution because sometimes it did not exist and
proposed an alternative solution to the noncooperative intergeneration game. As ex-
plained by Kydland (I 975b), in the language of dynamic games, Pollak used the feedback
solution and Peleg and Yaari the open-loop solution. For policy selection, the policymaker
is dominant, and for dominant-player games, the open-loop solution is inconsistent (see
Kydland 1975a, 1975b for further details). That is why Peleg and Yaari's solution was
not considered here.
X= Ext.
Xt
o consistent
quiI brium
0o opti al ut-u*
equilibrium
S(x1, Us).
best, given the current situation. The indifference curves imply that the
socially preferred inflation rate is zero, which seems consistent with the
public's preferences. We of course recognize that inflation is a tax on
reserves and currency, and a more informed public might prefer some
positive or negative inflation rate. If so, x, need only be interpre
as deviation from the optimal rate. The outcome of a consistent policy
selection clearly is not optimal. If the policymakers were compelled to
maintain price stability and did not have discretionary powers, the re-
sulting equilibrium would have no higher unemployment than the con-
sistent policy. The optimal equilibrium is point 0, which lies on a higher
indifference curve than the consistent-equilibrium point C.
It is perhaps worthwhile to relate our analysis to that of Taylor's
(1975), in which he found that the optimal monetary policy was random
in a rational-expectations world. Similar results would hold for our prob-
lem if uncertainty in the social objective function had been introduced.
Both for his structure and for ours, the optimal policy is inconsistent, and
consequently it is not optimal for the policymaker to continue with his
original policy rules.
and the objective is to minimize its expected value, the optimal value
for 2Zt will depend upon both yt and f~f, the policy rule which
used in the future. In other words, the best policy rule for the current
period frC(y) is functionally related to the policy rule used in the future
I-f(y), say
I-C = g(fif).
4 This is the solution concept used by Phelps and Pollak (1968) for an infinite-period
second-best growth problem when different generations had inconsistent preferences.
5 Optimal policy refers to the best policy, assuming it exists, within a certain class of
policies. Within the class of linear feedback rules fl(y,), we found that the best policy
rule depended upon the initial condition. The most general class of decision policies
are characterized by a sequence of probability measures indexed by the history
flt(xt, 7tyt) }, with the superscripted variables denoting all previously ob
of the variables. It was necessary to consider probability distributions because for some
games a randomized strategy will be optimal and not dominated by a deterministic one.
For games against nature, only deterministic strategies need be considered.
kt+2 = Xt + (1 - )kt+l,
where 3 is the constant physical rate of depreciation. Investment costs
associated with xt, the new investment plans in period t, occur in both
period t and period t + 1. This reflects the fact that time is required
to expand capacity, and investment expenditures occur over the entire
time interval. The fraction of the investment effort induced by plan xt
in the current period is 4, and the fraction induced in the subsequent
period is 1 - I . The investment rate in period t is then
Zt = qxt + (1 - O)Xt-1-
Following Haavelmo (1960), Eisner and Strotz (1963), Lucas (1967),
Gould (1968), and Treadway (1969), we assume that the investment
expenditures are an increasing convex function of the rate of capital
expansion zt. In order to insure constant returns to scale in the long
run, the function is assumed to have slope equal to the price q of capital
goods at zt = bkt. Making the quadratic approximation, the investment
expenditures in period t are then
In period t a tax rate 0 is applied to sales less labor costs and depreciation.
Letting T be the fraction of the "true" depreciation being tax deductible,
the tax bill is then
The view is that the adjustment cost term reflects costs internal to the
firm and therefore not eligible for the investment tax credit.6
The net cash inflow in period t is (8) less (9) plus (10). The objective
of the firm is to maximize the expected present value of this net cash
inflow stream. Maximizing each period's cash flow over the period's nt,
the objective function to be maximized becomes
where i = [1/(I - ))] [A(I - ox)] is output per unit of capital and f
is the discount factor.
The inverse aggregate demand function is assumed to be linear. Letting
capital letters denote the aggregates of the corresponding variables for
the individual firms, the inverse demand function is of the form
Pt = at- b&`K ,
where b is a positive constant, at is a stochastic demand shift parameter,
and Kt is the aggregate capital stock for the firms. We assume that at is
subject to the first-order autoregressive process
6 We are also implicitly assuming a per unit rather than a percentage tax credit.
7 If policy does not depend upon agents' decisions, the competitive equilibrium is
efficient and therefore maximizes the utility of the economy-wide consumer, a fact
exploited in Lucas and Prescott (1971) to characterize the competitive equilibrium.
For these examples, policy rules are of the feedback variety and depend upon past
decisions of private economic agents. In effect this introduces an externality. Suppose,
e.g., that future investment tax credits depend positively upon the magnitude of future
capital stocks. If all agents invest less now, future capital stocks will be smaller and,
consequently, future investment tax credits larger. Because of this externality, the com-
petitive equilibrium will not in general maximize the utility of the economy-wide
consumer, given the policy rule. This is why it was necessary to devise direct methods.
9 The values used for the fixed parameters were 3 = 0.1, 0 = 0.5, A = 1.15, y =
0.03, a = 0.28, b = 0.4, A = 0.7, q = 1, fl = ,B, = 0.9, 0 = 0.3, and OE = 0.03.
Given p, the parameter ui was chosen so as to give a mean of 2.5 for at. For instance, if
p = 0.6, as in the first example, we get p = 1. In the examples discussed, we used
g= 2.008, g2 = 0.2837, and g3 = 2.292, which are the stationary levels of the corre-
sponding target variables when the passive policy of 7rt = 0 is used in every period.
10 More details of the results of the numerical examples can be found in our original
working paper, which is available upon request.
VI. Discussion
We have argued that control theory is not the appropriate tool for
dynamic economic planning. It is not the appropriate tool because cur-
rent decisions of economic agents depend upon expected future policy,
and these expectations are not invariant to the plans selected. We have
shown that, if in each period the policy decision selected is the one which
maximizes the sum of the value of current outcomes and the discounted
valuation of the end-of-period state, the policy selected will be consistent
but not optimal. This point is demonstrated for an investment-tax-credit
policy example, using a rational-expectations equilibrium theory with
costs of adjustment and distributed lags for expenditures. In fact, active
stabilization effects did, for some distributed lag expenditure schedules,
contribute to economic instability and even make a stable economy
unstable.
Appendix
Lety be the state variables and x the decision variables for the firm. There is a
linear relationship between the next period's state variables,yt+ 1, and the cur
Xt andyt:
Yt + = f (Xt, Yt)* (Al)
The movement of economy-wide state variables Y and aggregate (or per firm)
decision variables X are described by the same linear function:
E fitR(xty
_t=0
Vt(yt, Yt, Wt) = R(xt,yt, Xt, Yt, Wt. 7rt) + flE[vt+ 1(Yt+1, Yt+1, Wt+1)] (A5)
subject to constraints (Al)-(A4) and one additional constraint. To explain this
last constraint, note that, since x, is chosen so as to maximize the valuatio
time t, if Vt is quadratic and the right-hand side of (A5) concave in x,, th
which maximizes the right-hand side of (A5), taking as given Xt, Yt, Wt, and t
motion of the economy-wide state variables, will be linearly related toyt, Xt, Y
Wt, and rt:
Xt = dt(ytX, Yt, J) Wt, 7t). (A6)
In order for the economy to be in equilibrium, we have to impose the constraint
that, when firms behave according to (A6), the aggregate or per firm Xt is ind
Xt. Therefore'1
Xt = dt (Yt, Xt, Yt, Wt, 7rt),
which can be rewritten as
given that policy is selected consistently and that the economy is competitive.
Thus the function ut gives the total expected value of the social objective functi
from period t throughout the rest of the horizon for the consistent policy. By
backward induction
u,(Yt, Wt) = it min {S(Xt, Yt, Wt, 7rt) + f&E[ut+ 1+(Yt+1, W+1 )]},
7t = It(Yt, Wt.).
yt = (kr, xt.-..),
and the linear equations governing its movement over time are
The equations governing the economy-wide variables are the same. Furthermore,
Pt = at- bAt
and
Zt = qxt + (1 - O)Xt-1-
Finally, the social objective function S given by
Equilibrium decision rules for agents were determined as the limit of first-period
decision rules as the life of the economy went to infinity. There is an interesting
and as yet unsolved problem as to the uniqueness of the equilibrium.1 2 For these
examples the equilibrium associated with a stationary policy rule did appear to
be unique, for when we used the method of successive approximation in the value
space (i.e., the v function in [A5]) the value function and therefore decision rules
converged to the same limit for a number of initial approximations. For some
unreasonable policy rules and finite T, there were no competitive equilibria.
Consistent solutions were computed in two different ways. The first determined
the first-period consistent policy for T-period problems and the limit determined
as T went to infinity. The second determined the nth approximation to the
consistent-equilibrium investment function X', given the nth approximation to
the consistent policy rule rln, using the methods described above. Optimal
control theory was then used to determine the policy flfn+1 which would be
optimal if X' were not to change as a result of the change in ITn. Given initial
linear feedback rule nol, sequences of linear rules {FIn, X" } were obtained. When
such sequences existed and converged, the limits constituted a consistent policy
rule and the corresponding equilibrium investment function. In no case did we
ever obtain two different consistent policies for the same structure, though both
methods of successive approximations were used and a number of different
starting values tried.
X = Gn(yY, W, 7r),
and given structural relations (A2) and (A3), one finds the optimal firm investment
function
x = d(y, X, Y, W 71),
X = DA(Y5 W, 7r)
Now let
References
Lucas, R. E., Jr. "Adjustment Costs and the Theory of Supply." J.P.E. 75, no.
4, pt. 1 (August 1967): 321-34.
. "Econometric Policy Evaluation: A Critique." In The Phillips Curve and
Labor Markets, edited by K. Brunner and A. H. Meltzer. Amsterdam: North-
Holland, 1976.
Lucas, R. E., Jr., and Prescott, E. C. "Investment under Uncertainty." Econo-
metrica 39 (September 1971): 659-81.
Muth, J. F. "Rational Expectations and the Theory of Price Movements."
Econometrica 29 (July 1961): 315-35.
Nordhaus, W. D. Invention Growth and Welfare: A Theoretical Treatment of Tech-
nological Change. Cambridge, Mass.: M.I.T. Press, 1969.
Peleg, B., and Yaari, M. E. "On the Existence of a Consistent Course of Action
When Tastes Are Changing." Rev. Econ. Studies 40 (July 1973): 391-401.
Phelps, E. S., and Pollak, R. A. "On Second-best National Saving and Game-
Equilibrium Growth." Rev. Econ. Studies 35 (April 1968): 185-99.
Phelps, E. S., and Taylor, J. B. "Stabilizing Properties of Monetary Policies under
Rational Price Expectations." Department of Economics Working Paper no.
75-7607, Columbia Univ., 1975.
Pollak, R. A. "Consistent Planning." Rev. Econ. Studies 35 (April 1968): 201-8.
Sargent, T. J. "Rational Expectations, the Real Rate of Interest, and the 'Natural'
Rate of Unemployment." Brookings Papers on Economic Activity, no. 2 (1973),
pp. 429-72.
Taylor, J. B. "Monetary Policy during a Transition to Rational Expectations."
J.P.E. 83, no. 5 (October 1975): 1009-22.
Treadway, A. B. "On Rational Entrepreneurial Behavior and the Demand for
Investment." Rev. Econ. Studies 36 (April 1969): 227-39.