1) This document examines optimal taxation and policy selection within a rational expectations framework using dynamic models with capital.
2) It finds that standard control theory techniques cannot be used to solve for the optimal taxation policy over time, as the optimal policy would be time inconsistent - what was optimal in the past would not be optimal now.
3) It develops a novel recursive method to overcome this difficulty by determining both the constraint set and value function recursively, even though there is doubt the optimal policy could actually be implemented due to its time inconsistency. The optimal policy still provides a benchmark to evaluate alternative time consistent policies.
Joao Guerreiro (Northwestern University), Sergio Rebelo (Northwestern University, NBER and CEPR), Pedro Teles (Católica-Lisbon School of Business & Economics, Banco
de Portugal and CEPR)
Joao Guerreiro (Northwestern University), Sergio Rebelo (Northwestern University, NBER and CEPR), Pedro Teles (Católica-Lisbon School of Business & Economics, Banco
de Portugal and CEPR)
Slides about non-existence of an equilibrium in DSGE models generally, how one can modify DSGE to evade this problem, and how an understanding of interest rate rule-based monetary policy should change. Describes how New Keynesian monetary policy (Taylor rule) works as well. Paper draft in https://papers.ssrn.com/abstract=3446931
Estimating Financial Frictions under LearningGRAPE
The paper studies the implication of initial beliefs and associated confidence under adaptive learning. We first illustrate how prior beliefs determine learning dynamics and the evolution of endogenous variables in a small DSGE model with credit-constrained agents, in which rational expectations are replaced by constant-gain adaptive learning. We then examine how discretionary experimenting with new macroeconomic policies is affected by expectations that agents have in relation to these policies. More specifically, we show that a newly introduced macro-prudential policy that aims at making leverage counter-cyclical can lead to substantial increase in fluctuations under learning, when the economy is hit by financial shocks, if beliefs reflect imperfect information about the policy experiment.
The dangers of policy experiments Initial beliefs under adaptive learningGRAPE
The paper studies the implication of initial beliefs and associated confidence on the system’s
dynamics under adaptive learning. We first illustrate how prior beliefs determine learning dynamics
and the evolution of endogenous variables in a small DSGE model with credit-constrained agents,
in which rational expectations are replaced by constant-gain adaptive learning. We then examine
how discretionary experimenting with new macroeconomic policies is affected by expectations that
agents have in relation to these policies. More specifically, we show that a newly introduced macroprudential policy that aims at making leverage counter-cyclical can lead to substantial increase in
fluctuations under learning, when the economy is hit by financial shocks, if beliefs reflect imperfect
information about the policy experiment. This is in the stark contrast to the effects of such policy
under rational expectations.
The heterogeneous effects of government spending, by Axelle Ferriere (Europea...ADEMU_Project
How expansionary is government spending? Evidence shows if
output increases, consumption doesn't decrease. Axelle Ferriere revisits this question, taking into account tax distribution.
Hierarchical Applied General Equilibrium (HAGE) ModelsVictor Zhorin
New techniques for uncertainty quantification (Chernoff entropy-based) in highly non-linear stochastic models; Tensor computing
Purpose: address the changing nature of service industries as providers of multi-dimensional contracts rather than simple price-based bundles of goods; models based on large data sets, consisting of country-wide surveys of
household data from Thailand, Mexico, Brazil, Spain
in combination with variety of geophysical and socioeconomic data. Innovative methods to represent complex data (Analytics/Visualization)
Purpose: evaluate household microfinance initiatives and credit expansion under different policies.
Agents gradually learn the structure of the economy.
Learning model delivers a sizeable recession in 2008-2010,
...whereas RE model predicts a counterfactual expansion.
In a medium scale model learning still matters.
This theory which is conveniently explain the concept of pricing through two independent Parameters which is DEMAND and SUPPLY.
This theory helps to determines the pricing of different factors of production like Land, Labour, Capital and Enterprenurs.
Research in behavioral economics has uncovered the widespread phenomenon of people making decisions against their own good intentions. In these situations, the government might want to intervene, indeed individuals might want the government to intervene, to induce behavior that is closer to what individual wish they were doing. The analysis of such corrective interventions, through taxes and subsidies, might be called ”behavioral public economics.” However, such analysis, where the government has an objective function that is different from that of individuals, is not new in public economics. In these cases the government is said to be ”non-welfarist” in its objectives, and there is a long tradition of non-welfarist welfare economics, especially the analysis of optimal taxation and subsidy policy where the outcomes of individual behavior are evaluated using a preference function different from the one that generated the outcomes. The object of this paper is to first of all present a unified view of the non-welfarist optimal taxation literature and, secondly, to present behavioral public economics as a natural special case of this general framework.
Slides about non-existence of an equilibrium in DSGE models generally, how one can modify DSGE to evade this problem, and how an understanding of interest rate rule-based monetary policy should change. Describes how New Keynesian monetary policy (Taylor rule) works as well. Paper draft in https://papers.ssrn.com/abstract=3446931
Estimating Financial Frictions under LearningGRAPE
The paper studies the implication of initial beliefs and associated confidence under adaptive learning. We first illustrate how prior beliefs determine learning dynamics and the evolution of endogenous variables in a small DSGE model with credit-constrained agents, in which rational expectations are replaced by constant-gain adaptive learning. We then examine how discretionary experimenting with new macroeconomic policies is affected by expectations that agents have in relation to these policies. More specifically, we show that a newly introduced macro-prudential policy that aims at making leverage counter-cyclical can lead to substantial increase in fluctuations under learning, when the economy is hit by financial shocks, if beliefs reflect imperfect information about the policy experiment.
The dangers of policy experiments Initial beliefs under adaptive learningGRAPE
The paper studies the implication of initial beliefs and associated confidence on the system’s
dynamics under adaptive learning. We first illustrate how prior beliefs determine learning dynamics
and the evolution of endogenous variables in a small DSGE model with credit-constrained agents,
in which rational expectations are replaced by constant-gain adaptive learning. We then examine
how discretionary experimenting with new macroeconomic policies is affected by expectations that
agents have in relation to these policies. More specifically, we show that a newly introduced macroprudential policy that aims at making leverage counter-cyclical can lead to substantial increase in
fluctuations under learning, when the economy is hit by financial shocks, if beliefs reflect imperfect
information about the policy experiment. This is in the stark contrast to the effects of such policy
under rational expectations.
The heterogeneous effects of government spending, by Axelle Ferriere (Europea...ADEMU_Project
How expansionary is government spending? Evidence shows if
output increases, consumption doesn't decrease. Axelle Ferriere revisits this question, taking into account tax distribution.
Hierarchical Applied General Equilibrium (HAGE) ModelsVictor Zhorin
New techniques for uncertainty quantification (Chernoff entropy-based) in highly non-linear stochastic models; Tensor computing
Purpose: address the changing nature of service industries as providers of multi-dimensional contracts rather than simple price-based bundles of goods; models based on large data sets, consisting of country-wide surveys of
household data from Thailand, Mexico, Brazil, Spain
in combination with variety of geophysical and socioeconomic data. Innovative methods to represent complex data (Analytics/Visualization)
Purpose: evaluate household microfinance initiatives and credit expansion under different policies.
Agents gradually learn the structure of the economy.
Learning model delivers a sizeable recession in 2008-2010,
...whereas RE model predicts a counterfactual expansion.
In a medium scale model learning still matters.
This theory which is conveniently explain the concept of pricing through two independent Parameters which is DEMAND and SUPPLY.
This theory helps to determines the pricing of different factors of production like Land, Labour, Capital and Enterprenurs.
Research in behavioral economics has uncovered the widespread phenomenon of people making decisions against their own good intentions. In these situations, the government might want to intervene, indeed individuals might want the government to intervene, to induce behavior that is closer to what individual wish they were doing. The analysis of such corrective interventions, through taxes and subsidies, might be called ”behavioral public economics.” However, such analysis, where the government has an objective function that is different from that of individuals, is not new in public economics. In these cases the government is said to be ”non-welfarist” in its objectives, and there is a long tradition of non-welfarist welfare economics, especially the analysis of optimal taxation and subsidy policy where the outcomes of individual behavior are evaluated using a preference function different from the one that generated the outcomes. The object of this paper is to first of all present a unified view of the non-welfarist optimal taxation literature and, secondly, to present behavioral public economics as a natural special case of this general framework.
The role of events in simulation modelingWeibull AS
The need for assessing the impact of events with binaryi outcomes, like loan defaults, occurrence of
recessions, passage of a special legislation, etc., or events that can be treated like binary events like
paradigm shifts in consumer habits, changes in competitor behavior or new innovations, arises often
in economics and other areas of decision making.
By using analogies from intervention analysis a number of interesting and important issues can be
analyzed:
If two events affects one response variable will the combined effect be less or greater than the sum of both?
Will one event affecting more than one response variable increase the effect dramatically?
Is there a risk of calculating the same cost twice?
If an event occurs at the end of a project, will it be prolonged? And what will the costs be?
Questions like this can never be analyzed when using a ‘second layer lump sum’ approach. Even
more important is possibility to incorporate the responses to exogenous events inside the simulation
model, thus having the responses at the correct point on the time line and by that a correct net
present value for costs, revenues and company or project value.
Econ 3022 MacroeconomicsSpring 2020Final Exam - Due A.docxtidwellveronique
Econ 3022: Macroeconomics
Spring 2020
Final Exam - Due April 24th 11:59pm
1 Multiple Choice Questions (5 points each)
Question 1 What is Ricardian Equivalence?
(a) The economic hypothesis that agents’ decisions are una↵ected by the timing of taxation
and government spending
(b) The economic hypothesis that agents’ decisions are a↵ected by the timing of taxation
and government spending
(c) The economic hypothesis that taxation must be equal every period.
(d) The economic hypothesis that it is impossible to individually identify taxation today
and taxation tomorrow.
Question 2 Consider the consumer problem from the microeconomic foundations we dis-
cussed in class. Suppose the wage decreases. What do we expect to happen to house-
hold labor supply?
(a) Unclear
(b) Increase
(c) Decrease
(d) Stay constant
1
Question 3 Consider the consumer problem from the real intertemporal model. Which of
the following conditions must be satisfied at the solution?
(a) MRSl,c = w
(b) MRSc0,l0 =
1
w0
(c) MRSl,l0 =
w(1+r)
w0
(d) All of the above
Question 4 If total factor productivity tomorrow, z0, increases. What should happen to
investment?
(a) Unclear
(b) Increase
(c) Decrease
(d) Stay constant
Question 5 Consider the standard Solow model from class where the production function
is zF (K, N) = zK↵N1�↵. What is the golden rule savings rate?
(a) sgr = 1 � ↵
(b) sgr = ↵
(c) The savings rate that leads to a steady state with the highest level of income per capita
(d) The savings rate that leads to a steady state with the lowest level of income per capita
2
2 Economic Growth (20 points)
Consider the Solow Growth Model seen in class where the production function is Cobb-
Douglas and given by:
Y = zK↵ (N)
1�↵
where 0 < ↵ < 1 and z is a constant. Let s be the savings rate of this economy, so that
aggregate savings is just a constant fraction of aggregate output: S = sY . Let n be the rate
of population growth, so N
0
N
= 1 + n. Finally, let d be the depreciation rate, and assume the
law of motion for aggregate capital is given by:
K
0 = (1 � d) K + I
(a) (5 pts) Find an expression for the steady state level of capital per capita (k⇤) that only
depends on parameters of the model. Clearly show your work.
(b) (5 pts) Discuss how per capita variables (consumption and income) as well as aggregate
variables (consumption, capital stock, output, and savings) behave in steady state.
Now, suppose that we have a linear production function given by
Y = zK
where z is a constant. Let s be the savings rate of this economy, so that aggregate savings
is just a constant fraction of aggregate output: S = sY . Let n be the rate of population
growth, so N
0
N
= 1 + n. Finally, let d be the depreciation rate, and assume the law of motion
for aggregate capital is given by:
K
0 = (1 � d) K + I
(c) (5 pts) Find an expression for the level of per capita capital stock today as a function
of per capita capital stock tomorrow. Clea.
American Economic Association Some International Evid.docxnettletondevon
American Economic Association
Some International Evidence on Output-Inflation Tradeoffs
Author(s): Robert E. Lucas, Jr.
Source: The American Economic Review, Vol. 63, No. 3 (Jun., 1973), pp. 326-334
Published by: American Economic Association
Stable URL: http://www.jstor.org/stable/1914364
Accessed: 08-02-2017 17:06 UTC
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 [email protected]
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms
American Economic Association is collaborating with JSTOR to digitize, preserve and extend access to The
American Economic Review
This content downloaded from 131.162.129.148 on Wed, 08 Feb 2017 17:06:22 UTC
All use subject to http://about.jstor.org/terms
Some International Evidence on
Output-Inflation Tradeofs
By ROBERT E. LUCAS, JR.*
This paper reports the results of an
empirical study of real output-inflation
tradeoffs, based on annual time-series from
eighteen countries over the years 1951-67.
These data are examined from the point
of view of the hypothesis that average
real output levels are invariant under
changes in the time pattern of the rate of
inflation, or that there exists a "natural
rate" of real output. That is, we are con-
cerned with the questions (i) does the
natural rate theory lead to expressions of
the output-inflation relationship which
perform satisfactorily in an econometric
sense for all, or most, of the countries in
the sample, (ii) what testable restrictions
does the theory impose on this relation-
ship, and (iii) are these restrictions con-
sistent with recent experience?
Since the term "'natural rate theory"
refers to varied aggregation of models and
verbal developments,' it may be helpful
to sketch the key elements of the particular
version used in this paper. The first
essential presumption is that nominal out-
put is determined on the aggregate demand
side of the economy, with the division
into real output and the price level largely
dependent on the behavior of suppliers of
labor and goods. The second is that the
partial "rigidities" which dominate short-
run supply behavior result from suppliers'
lack of information on some of the prices
relevant to their decisions. The third
presumption is that inferences on these
relevant, unobserved prices are made
optimally (or "rationally") in light of the
stochastic character of the economy.
As I have argued elsewhere (1972),
theories developed along these lines will
not place testable restrictions on the co-
efficients of estimated Phillips curves or
other single equation expressions of the
tradeoff. They will not, for examp.
Dear students get fully solved assignments
Send your semester & Specialization name to our mail id :
help.mbaassignments@gmail.com
or
call us at : 08263069601
The presentation aims to explain the meaning of ECONOMETRICS and why this subject is studied as a separate discipline.
The reference is based on the book "BASIC ECONOMETRICS" by Damodar N. Gujarati.
For further explanation, check out the youtube link:
https://youtu.be/S3SUDiVpUGU
The Art Pastor's Guide to Sabbath | Steve ThomasonSteve Thomason
What is the purpose of the Sabbath Law in the Torah. It is interesting to compare how the context of the law shifts from Exodus to Deuteronomy. Who gets to rest, and why?
How to Create Map Views in the Odoo 17 ERPCeline George
The map views are useful for providing a geographical representation of data. They allow users to visualize and analyze the data in a more intuitive manner.
We all have good and bad thoughts from time to time and situation to situation. We are bombarded daily with spiraling thoughts(both negative and positive) creating all-consuming feel , making us difficult to manage with associated suffering. Good thoughts are like our Mob Signal (Positive thought) amidst noise(negative thought) in the atmosphere. Negative thoughts like noise outweigh positive thoughts. These thoughts often create unwanted confusion, trouble, stress and frustration in our mind as well as chaos in our physical world. Negative thoughts are also known as “distorted thinking”.
Model Attribute Check Company Auto PropertyCeline George
In Odoo, the multi-company feature allows you to manage multiple companies within a single Odoo database instance. Each company can have its own configurations while still sharing common resources such as products, customers, and suppliers.
The Indian economy is classified into different sectors to simplify the analysis and understanding of economic activities. For Class 10, it's essential to grasp the sectors of the Indian economy, understand their characteristics, and recognize their importance. This guide will provide detailed notes on the Sectors of the Indian Economy Class 10, using specific long-tail keywords to enhance comprehension.
For more information, visit-www.vavaclasses.com
Sectors of the Indian Economy - Class 10 Study Notes pdf
Dynamic optimal taxation,
1. Journal of Economic Dynamics and Control 2 (1980) 79-91. 0 North-Holland
.
DYNAMIC OPTIMAL TAXATION,
RATIONAL EXPECTATIONS AND OPTIMAL CONTROL
Finn E. KYDLAND and Edward C. PRESCOTT*
Carnegie-Mellon University, Pittsburgh, PA 1.5213, USA
Received August 1979
Within a rational expectations framework, policy has effect if it alters relative prices and policy
evaluations are exercises in modern public finance theory. The time inconsistency of an optimal
taxation plan precludes the use of standard control theory for its determination. In this article
recursive methods are developed that overcome this difficulty. The technique is novel in that the
constraint set as well as the value function are determined recursively. Even though there is little
hope of the optimal plan being implemented - because of its time inconsistency - we think the
exercise is of more than pedagogical interest. The optimal plan’s return is a benchmark with
which to compare the time consistent solution under alternative institutional constraints which
society might choose to impose upon itself.
1. Introduction
The purpose of this article is to examine the problem of optimal policy
selection within the rational expectations competitive framework. The natural
context within which to discuss optimal policy selections is optimal tax-
ation.’ This allows us to be fairly explicit as to how the policymaker’s
objective is related to those of the economic agents. We assume that there is
a representative consumer, and that there is no conflict between public and
private objectives. The individual values public goods, but prefers not to pay
taxes. The role of the policymaker is to provide public goods in such a way
as to maximize the utility of the representative consumer subject to the
constraint that goods are financed by proportional taxes upon labor and
capital incomes.’
We first present a standard static taxation model, for which optimization is
fairly straightforward. We then consider a multiple period problem with
*We thank Robert E. Lucas, Jr., Agnar Sandmo, John B. Taylor, Henry Wan, and
participants of the 1979 Warwick Summer Workshop for helpful comments. We acknowledge
the financial support of the National Science Foundation.
‘For a review of the optimal taxation literature, see Sandmo (1976).
‘Justifications for ruling out non-distortionary taxes of the lump sum variety are based upon
heterogeneity of consumers, which is not an element of our models. Heterogeneity was not
introduced because, even in static models, it greatly complicates the analysis [cf. Romer (1975)]
and would obscure the points we wished to make.
2. 80 F.E. Kydland and E.C. Prescott, Dynamic optimal taxation
capital and show that standard control-theory techniques are not applicable
to the optimal taxation programming problem. 3 Bellman’s (1957) principle of
optimality fails and an optimal policy will in general be time inconsistent.
That is, the optimal policy plan at time t’ greater than t is not the
continuation of the plan that was optimal at time t.4 Not only does this
severely complicate the determination of the optimal policy by precluding the
applications of optimal control theory, but more importantly, the time
inconsistency makes it doubtful whether such a policy would ever be
implemented. Rather we would expect the time consistent solutions to be
adopted, that is, the rule resulting if the policy action taken at each point in
time is best taking into consideration both the current period outcome and a
correct evaluation of the end of period position. Unlike games against
nature, for which control theory was developed, the consistent solution is in
general not optimal and as shown in Kydland and Prescott (1977) may be
very suboptimal.
In spite of the time inconsistency problem, we do not think the de-
termination of the optimal policy and the resulting return is without interest.
The optimal return is a benchmark with which to compare the return of the
time consistent policy under a particular set of institutional constraints.
Possibly a constitutional amendment requiring the budget to be balanced in
peacetime or an institutional arrangement which results in the process of
policy change being long and protracted might result in the time consistent
solution being nearly optimal. If so, it would provide strong justification for
society establishing such arrangements.
2. A simple static optimal taxation model
. Consider an economy with a large number of small economic agents.
Assume that the preferences of the representative household can be repre-
sented by a utility function u(c, n, g), where c is consumption, n is labor
supply, and g is public expenditures. With a representative household, the
social welfare function is well defined, namely to maximize the same utility
function. The decision variables of the household are c and n, and, assuming
a linear production function y=on, the household is constrained by
Oscs(l-r)on and OSnSfi,
‘An excellent overview of control theory with emphasis on its applicability to economics can
be found in Chow (1975). Holly and Zarrop (1979), using non-recursive methods, have
developed and applied an algorithm for computing optimal solutions for finite horizon,
deterministic problems.
“The point was established in Kydland and Prescott (1977) and discussed at greater length in
Prescott (1977). Calvo (1978) has demonstrated the time inconsistency of an optimal monetary
policy. As inflation can be viewed as a tax on liquidity, his results can be viewed as being within
the optimal taxation framework and therefore as being complementary to this analysis.
3. F.E. Kydland and E.C. Prescott, Dynamic optimal taxation 81
where T is the income tax rate and ii is the maximum amount of labor
services that can be supplied. Note that w is the marginal product pf labor
and therefore in equilibrium the wage.
Assume now that z and g are set by a policymaker with the objective of
maximizing the utility of the representative consumer, i.e.,
max u(c,n,g1. (1)
where the aggregate variables are appropriately measured in per capita
terms. One constraint is that there be sufficient revenue to finance govern-
ment purchases; that is,
zwn 2 g.
The tax rate is constrained by OsT si< 1.
The second set of constraints come from the maximization problem of the
stand-in consumer. In order to insure a unique interior solution, we make
the following fairly standard assumptions: The utility function is strictly
increasing and strictly concave in consumption c and the negative of labor
supplied -n and is continuously differentiable. In addition
for all 0s T s i and 05 g6 iiw, where u, and u, are the partial derivatives
with respect to c and n, respectively. The consumer maximizes his utility over
his decision variables c and n subject to his budget constraint which will be
binding. Thus, his problem is
max u(c,4 gh
E.”
subject to
c= (1 -z)on.
With the above assumptions, the consumer is in equilibrium if and only if
u,k,n,g)= -(l-Tbq(c,n,g). (3)
Thus, the policy problem is to maximize (1) subject to constraints (2) and (3),
the inequality constraints
T ji, n 5 ii, c+g~wn,
4. 82 F.E. Kydland and E.C. Prescott, Dynamic optimal taxation
and the non-negativity constraints. This is a well-defined programming
problem. The compactness of the constraint set and continuity of the
objective function guarantee the existence of a solution.
3. The dynamic certainty case
We now introduce capital which makes the problem dynamic. We shall
consider the simplest type of model in which consumers own capital which
they rent to firms. The firms use capital and labor as inputs to produce
output which can either be consumed in the same period, be used as a public
good, or be used to augment the capital stock. The decision variables of the
consumer in period t are consumption, labor supply, and how much capital
to carry over to the next period. Finally, public good expenditures are
financed by taxes on labor and capital income.
For this problem we can define an equilibrium where the prices clear the
markets for given present and future values of the government policy
variables. Using the price of output as a numeraire, these equilibrium prices
would be {w:}PO_~ and {r:}z,, where w, is the wage rate and rl the rental
rate of capital in period t. In what follows, we shall omit the stars and
assume that we are referring to the equilibrium prices in stating the
constraints imposed by the rational expectations maximizing assumptions.
We assume that the utility function of the representative consumer is time
separable with discounting:
Consumer’s Problem:
max f pIu(c,, 4, g,),
t=o
subject to
k,+,+c,~k,+(l-e,)r,k,+(1-7,)w,n,,
ct,n,,k,+,20, n,Sfi, t=O , 1,..., k, given.
Here 0 5 0,s 8~ 1 and 0 5 7,s T-C 1 are the tax rates applicable to capital and
labor income, respectively. _
Factor supply and product demand are determined from the following
first-order conditions for t =0, 1, . . . :’
u,k,n,, g,)= - U-7,bv,(c 1,nn gI )7 (4)
‘Assumptions similar to those of the previous section concerning the function u(c,n,g) are
easily developed that guarantee an interior solution for consumption and labor supply with
uniformly bounded marginal utility for all feasible tax policies.
5. F.E. Kydland and E.C. Prescott, Dynamic optimal taxation 83
(5)
.
Since the consumer owns the capital, the firm’s problem is static:
Firm’s Problem:
maxCf(k,,n,)-r,k,-w,nJ, t=O,l,...,
where f(k, n) is a constant-returns-to-scale production function. In equilib-
rium, profits are zero and therefore they need not be included in the budget
constraint of the consumer. The profit-maximizing conditions are simply
(6)
and
(7)
We assume that f(0, 1 )= w > 0 and that f(k, n) is strictly concave in k, strictly
increasing in both its arguments, and positive. These assumptions guarantee
that in equilibrium (6) and (7) must be satisfied. We also assume there is a E
such that L=f(&ti) and that k, s E. Element f; is the maximal sustainable
capital stock.
The optimal taxation problem is then to choose {TI,=(g,, Ot,~l)}zo so as to
maximize
,$lo (c,,4, g,h
P’u
subject to
gr14r,k,+r,wn,9 t=O,l,...,
and constraints (4~(7), which are imposed by the rational expectations
equilibrium assumption.
The labor supply-consumption decisions, n, and c,, of the consumer depend
not only upon the current state of the economy, k,, and current tax rates,
(e,,t,), but also upon future tax rates. Until the sequences offuture tax rates
are specified the equilibrium current decisions of the consumers cannot be
determined.
Suppose the optimal policy sequence {np = (gp,@,$)}p”,, exists and is
unique. This optimal policy will be time inconsistent in the sense that at time
s>O the policy {$}lZs will not be optimal at time s. The reason it is not
optimal is because current equilibrium decisions of the consumer are
6. 84 I~‘.E. Kydland and E.C. Prescott, Dynamic optimal taxation
functions of the current state, current policy decisions, and anticipated future
policy actions.
Before proceeding further, a definition of state variable is needed.
3.1, Definition of state variable
A decision problem is Markovian if after any number of periods, say t, the
effect of decisions of the current and subsequent periods upon the total
return depends only upon the state of the system at the beginning of the tth
period and subsequent decisions. A preference-information-technology-wealth
distribution structure is Markovian if each agent’s decision problem is
Markovian in the assumed state variable vector, given that other agents’
decision problems are Markovian in that state vector.
The state variable must, among other things, reflect the effect of past
decisions upon the subsequent production possibility set. This is typically
accomplished by including capital stocks and inventory stock variables as
components of the state vector. When preferences are time separable, no
state variables are needed to index the effect of past decisions upon the
individual’s ordering of subsequent consumption paths (or distributions of
subsequent consumption paths if there is uncertainty). For some analyses, the
current distribution of money, bonds, and stocks must be specified by the
state vector. Knowing the current state is sufficient for determining the
current and subsequent competitive allocations and equilibrium prices (or
process governing them if there is uncertainty). With this definition, prices
and current-period decision variables are not state variables. Current equilib-
rium prices and decisions are functions of the state variables.
The naive application of optimal control is likely to lead to a consistent
solution for reasons given in Kydland and Prescott (1977). The solution is
consistent in the sense that it is best, given the current state variables and
that policies will be similarly selected in the future. This consistent solution
will be suboptimal, however, because the effects of the policy for any future
period t on agents’ behavior in earlier periods are not taken into account.
For the optimal taxation example the consistent solution for the current
period is to first tax away all the capital income because the capital is
already there and does not enter directly into the utility function, or, in other
words, it is supplied inelastically. Anticipating this, agents will save little and
the capital stock will be small.6
A superior alternative would be to restrict policymakers to use a well-
understood policy rule with good operating characteristics. Even if the best
one among such rules could be determined, however, the time inconsistency
61n his paper on optimal taxation, Ramsey (1927, p. 59) brielly considers the dynamic
problem, but because it is ‘considerably more dilkult essentially assumes the dynamics away.
7. F.E. Kydland and E.C. Prescott, Dynamic optimal taxation 85
of this policy makes it a doubtful question whether policymakers would
continue to use it in future periods. If not, expectations of economic agents
would clearly be affected, leading to a change in their behavior. It is still of
interest, however, to study optimal policy over time, even though it is time
inconsistent. For example, when a new tax system is introduced, this is often
a relevant restriction; more generally, one might imagine large costs as-
sociated with changes in policy.
3.2. Determining optimal policy
The time inconsistency severely complicates the computation of the
optimal policy. Standard recursive methods are no longer applicable. In what
follows we outline a possible computational procedure, and point out the
difficulties involved.
To obtain restrictions imposed by the rational expectations equilibrium
assumption, we formulate the Lagrangean for the consumer
L= f plIU(cI1nt,gr)+i,[(l-r,)w,n,+(1-8,)r,k,+k,-k,+,-c,]}.
r=o
The first-order conditions are’
u,k,, n,,g,)= 4, (8)
4 Cc,, g,)+ 4 (1- T,)w,= 0,
4, (9)
BC1+(l-~,+,)r,+Il~r+l-~,=O, t=O, l,... . (10)
In addition, we have the profit-maximizing conditions (6) and (7).
Using (6~(9) along with the budget constraint of the consumer, we can
write
x, = d(4, n,,4 ), t=O, l,..., (11)
where x, = (c,, n,, k, + 1) are the decision variables of the.consumer. Constraint
(lo), using (6), becomes
(12)
But, from (ll), n,,, is a function .. of kt+,, 7~,+~ and A,+lr and (12) can
‘These equations are equivalent to eqs. (4) and (S), and, in fact, represent a derivation of (4)
and (5).
8. 86 F.E. Kydland and E.C. Prescott, Dynamic optimal taxation
therefore be written as
4=hh,k,4-1), t = 1,2,. . .) (13)
This constraint is unusual in that it goes backwards in time. It says that the
ratio of next period’s to this period’s marginal utilities with respect to
consumption must equal the after-tax rate of transformation between con-
sumer goods next period and consumer goods this period. The introduction
of a pseudo-state variable A,- i, which is a shadow price, is necessary because
of the time inconsistency of the optimal policy.
The problem is not a Markov decision problem [see Bellman (1957)]
because there is another constraint which must be determined recursively.
Let !G!be the set of (k,, A,- i) for which there exists a policy sequence with an
equilibrium. That is, there exists sequence {xS,7c,,A,}~=, such that constraints
(11) and (13) are satisfied for periods s 2 t, where t > 0. For A,_ i sufficiently
small, and therefore consumption in the previous period sufficiently high, no
policy sequence for which there is an equilibrium will exist. This necessitates
the addition of constraints
(4+1,J,)EQ, t=0,1,2 )... .
We define the elements ~min and A,,, to be the minimum and maximum,
respectively, of I.+(c, n, g) over the set of (c, n, g) for which c, n, g 20,
c+gSf(l;,n), and nsii.
Dropping the time subscript and using the prime to denote next period
value and the minus one subscript last period value, we define the following
mapping @ of closed subsets of Zm [0, k] x [n,i”, A,,J into the same space:
@(Q)=((k,L,)EZ: there exists (71,x, A)
satisfying the constraints below},
x=(c,n,k’)=d(k,7c,i),
.
A=h(7t,k,L1),
(k’,l)EL?.
In addition there are the non-negativity constraints and maximum con-
straints on tax rates and labor supplied. We seek the largest set which is a
fixed point of @. The mapping @ is monotonic in the sense that Sz’G Sz
implies Q(U) G @(sZ).Consider the decreasing sequence of sets
Qi+ 1= @(12~) where Sz,= Z.
The limit of this sequence is a greatest fixed point of @, and is the recursively
9. F.E. Kydland and E.C. Prescott, Dynamic optimal taxation 87
determined constraint set 52. This set is non-empty, for given any k E [0, E],
there is an equilibrium path for policy 0, = r, = g, = 0. This implies in addition
that the projection of D on [0, iZ] is [0, 121.
Let u(k,, 1, _ r ) be the maximum present value at time t over all feasible
current and future policies given the constraints implied by the rational
expectations equilibrium concept for current and future private decisions,
and given the current capital stock k, and last period’s marginal utility A,- r,
which provides the link to the past. Then part of the policy problem is to
solve the functional equation
u(k,~-,)=maxCu(c,n,g)+pu(k’,1)],
n,*,.a
subject to the constraints
x=(c,n, k’)=d(k,n,A),
I=h(n, k,L,),
gIefk(k,n)k+~S.(k.nh
x,?T~O,
7 Si, tlsB
and
(k’,A)EQ where (k,I-,)ES2.
The optimal policy is of the form n=n(k, A- r). This means that at time t,
given k, and A,- r (which define admissible combinations of n, and A,), we
obtain II, = n (k,, A, _ r ), which determines Ar, and therefore c,, n,, and k, + 1.
Note that constraint (13) puts no restrictions on A,. Given k,, the
policymaker can choose rrO, x0, and 1, which maximize
u(co,no,g,)+Bu(k,,I,),
subject to
xo = d (ko, no, A, ),
go~~ofk(ko~~o)ko+70fn(kO~~o)~o~
(k,,lo)EQ.
This final part of the policy problem of choosing the initial I, determines the
optimal policy for the entire future. Thus the pseudo-state variable I,-, is a
10. 88 F.E. K.~~//trd cd EC. Prescott, Dynumic optimcrl taxation
device to ensure that the effi?cts offuture policies on agents’ behavior in earlier
periods are taken into account.’
4. An extension to uncertainty
In this section we outline- the extension of our analysis to environments
with uncertainty. Suppose preferences and technology are affected by random
shocks. For example, some exogenous random event may affect the demand
for national defense, or a new scientific discovery may affect the production
possibility set. In this section we assume the utility function of the repre-
sentative individual is state contingent depending upon the shock s. For
example, the value of national defense may vary over time depending upon
the world political situation. The consumer maximizes
subject to his budget constraint. From his point of view, government
expenditure g, and the tax-rate parameters (e,,r,), which affect his budget
constraint, are exogenous as is the shock s,. For simplicity, the process
governing s, is assumed to be a Markov chain with m possible states. The
transition probabilities PrGli) are the probabilities that s,, i =i given s,=i
for i,i=l,..., m.
To simplify notation, the time subscript has been omitted in the sub-
sequent development. Prime variables denote next-period variables, and the
subscript - 1 last period. Let v(s-i, k,1-,) be the maximal obtainable
expected utility for the representative individual given that the beginning-of-
period capital stock was k and that A-i and the consumer equilibrium
decisions this period are consistent.
The proposed dynamic program is
v(s-,,kL1)=max f Pr(sIs-,)Cu(c,,n,,g,,s)+pv(s,k6,~,)1, (14)
s= 1
subject to constraints for s = 1,. . ., m,
kl+c,=k+(l-r,)f,(k,n,)+(l-e,)f,(k,n,),
(1-M, (k n,)u,(c,,n,,g,,s)= -u, k, n,,g,,s),
1, = 4 k, n,,g,,s1,
tsn,f,(k,n,)+e,kf,(k,n,)Lg,,
(Kk)ENs,,
‘This argument was partly motivated by comments in BryGnt (1977).
11. F.E. Kydland and E.C. Prescott, Dynamic optimal taxation 89
and constraint
.
n-,=P 5 Pr(s(s-,)u,(c,,n,,g,,s)Cl+(l-e,)f,(k,n,)l. (15)
s=I
The decision variables are individual decision X = {c,, n,, kj},“= i, policy de-
cision l7 = {g,, Os,t,}~Z i and n = {A,}:= r. The constraints that (&,I,) belong
to set Q(s) are feasibility constraints. The sets Q(s) must be such that there is
a solution to the right-hand side of (14) if and only if (k, A- i)~ Q(s- i) for
s-1=1,..., d. Thus, like the value functions u, the constraint sets are
determined recursively. One can, at least in theory, find a solution which is
computable as the limit of finite T-period problems as T becomes infinite.
The optimality eq. (14) is of interest for the maximization over elements
specifying actions contingent upon s - that is, prior to observing s.
Consequently, s is not a state variable in the sense of dynamic programming.
This is necessary because of the form of constraint (15). It jointly depends
upon decisions contingent upon all s and not just the realized s.
As for the deterministic case, constraint (15) is not applicable in the initial
period and choosing I for period zero is part of the optimization problem.
This choice determines the optimal shock-contingent policy for the entire
future. Let the optimal decision rules be
The actual policy selected depends upon the realized shock. If the shock is s,
the policy action is the sth component of l7, and the equilibrium decision of
the representative consumer the sth component of X. The state variables for
the subsequent period are the realized shock s, the consumer’s equilibrium k’,
and the sth component of A.
To summarize, in the abstract at least, even the stochastic case can be
formulated as a recursive problem using the pseudo-state variable 1-i along
with recursively determined constraint sets G(s). This formulation leads to
unusual constraints, however, and the problem of actually computing an
optimal policy would appear quite formidable even for relatively simple
parametric structures.
5. Concluding remarks
We have not argued that optimization and quantitative methods should
not be used for policy evaluation and selection within self-fulfilling and
rational expectations environments. Rather we have argued that naive
12. 90 F.E. Kydland and E.C. Prescott, Dynamic optimal taxation
application of optimal control theory methods will result in time-consistent
but suboptimal policies. As was demonstrated in Kydland and Prescott
(1977) and Prescott (1977), the consistent solution can be very suboptimal.
We think the determination of optimal policy, and more importantly its
return, is important. It provides a standard with which to compare time
consistent policy under alternative institutional constraints. In environments
where policy rules can be changed only after an extended period of public
deliberation, the time consistent rule may be nearly optimal.
We do not think discounting is the crux of the consistency problem. For
our optimal taxation problem, the results were insensitive to the discount
factor p and the inconsistency problem did not disappear as /I approached
one. If j equals one, there are problems with existence of competitive
equilibria because the infinitely-lived representative household’s utility is
infmite. We suspect this problem can be circumvented by introducing a
Ramsey ‘deviation from bliss’ preference ordering. With such an analysis the
inconsistency problem will remain.
Finally, we emphasize that the fixed-rule procedure we advocate does not
necessarily imply constant values or constant growth rates of the policy
instruments. Quite possibly a feedback rule with the tax parameters varying
systematically with economic conditions may dominate any policy of con-
stant tax parameters. A policy rule, however, is needed before one can
predict what equilibrium process will govern the economy and an impli-
cation of dynamic maximization theory is that a policy-rule invariant
autoregressive model will not exist.
It should be noted that our equilibrium framework led to decision rules
that depended not only on the state and tax rates, but also on the
unobservable shadow price 2. This causes an estimation problem. To the
extent, however, that everything in eq. (12), except 1, and 1,+ i, can be
observed, including the discount factor /I, at least the ratios of next period’s
to this period’s shadow price could be determined, and these prices can
clearly be scaled any way we want: Such relationships might therefore
conceivably be estimated.g But even if the relationships can be identified and
estimated, we have not yet found practical methods for computing the
optimal policy when there are stochastic constraints because of the high
dimensionality of the dynamic program.
‘Identitied, policy-invariant structures specifying current decisions as functions of current
state, current prices, and expectations of future prices and policy parameters exist. Huntzinger
(1979) and Taylor (1979) have estimated such structures and Wallis (1979) has developed the
econometrics much further. An alternative approach is that of Hansen and Sargent (1980) who
use maximum likelihood methods to estimate parameters of preferences, technology, and
exogenous processes directly.
13. FL Kydland and EC. Prescott, Dynamic optimal tuxution 91
Bellman, Richard E., 1957, Dynamic programming (Princeton University Press, Princetom NJ).
Bryant, John, 1977, Should control theory be used for economic stabilization?: A comment, in:
K. Brunner and A.H. Meltzer, eds. Optimal policies, control theory and technology exports
(North-Holland, Amsterdam) 99-100.
Calve, Guillermo A., 1978, On the time consistency of optimal policy in a monetary economy,
Econometrica 46, 1411-1428.
Chow, Gregory C., 1975, Analysis and control of dynamic economic systems (Wiley, New York).
Hansen, Lars P. and Thomas J. Sargent, 1980, Formulating and estimating dynamic linear
rational expectations models, Journal of Economic Dynamics and Control. This issue.
Holly, S. and M.B. Zarrop, 1979, Calculating optimal economic policies when expectations are
rational, PROPE Discussion Paper no. 33 (Imperial College of Science and Technology,
London).
Huntzinger, R. LaVar, 1979, Market analysis with rational expectations: Theory and estimation,
Journal of Econometrics 10, 127-145.
Kydland, Finn E. and Edward C. Prescott, 1977, Rules rather than discretion: The inconsistency
of optimal plans, Journal of Political Economy 85, 473491.
Prescott, Edward C., 1977, Should control theory be used for economic stabilization?, in: K.
Brunner and A.H. Meltzer, eds., Optimal policies, control theory and technology exports
(North-Holland, Amsterdam) 13-38.
Ramsey, Frank P., 1927, A contribution to the theory of taxation, Economic Journal 37, 47-61.
Romer, Thomas, 1975, Individual welfare, majority voting, and the properties of a linear income
tax, Journal of Public Economics 4, 163-185.
Sandmo, Agnar, 1976, Optimal taxation: An introduction to the literature, Journal of Public
Economics 6, 37-54.
Taylor, John B., 1979, Estimation and control of a macroeconomic model with rational
expectations, Econometrica 47, 1267-1286.
Wallis, Kenneth F., 1979, Econometric implications of the rational expectations hypothesis,
Working Paper (Warwick University, Coventry). Revised version forthcoming in:
Econometrica.