1
si
>CHU^
^
*
5 (iSBSAS^j §
Digitized by the Internet Archive
in
2011 with funding from
Boston Library Consortium
Member
Libraries
http://www.archive.org/details/onillsofadjustmeOOcaba
DEWEY J
working paper
department
of economics
ON THE ILLS OF ADJUSTMENT
Richardo
J.
Mohamad
95-21
Caballero
L.
Hammour
Jul.
1995
massachusetts
institute of
technology
50 memorial drive
cam bridge, mass. 02139
ON THE ILLS OF ADJUSTMENT
Richardo
J.
Mohamad
95-21
Caballero
L.
Hammour
Jul.
1995
MASSACHUSETTS INSTITUTE
NOV
1 4 1995
UDP«?^iuO
95-21
On
Ricardo
the
of
Ills
Mohamad
Caballero
J.
Adjustment
This version: July
17,
L.
Hammoui*
1995
Abstract
We
analyze market impediments to the process of structural adjustment.
on incomplete-contract
inefficiencies in the transactions
We
focus
between workers and firms that
render the quasi-rents from "specific" investment appropriable. During adjustment, the
result
is
a depressed rate of creation of the new productive structure and excessive de-
struction of the old one, leading to an
weakens the incentives
tive structure.
An
employment
for extensive restructuring,
crisis.
and
Moreover, appropriability
results in
a
contracting one. In contrast, the
common
effective "synchronizer" of creation
also reducing
produc-
adequate managed-adjustment program ought to combine vigorous
creation incentives in the expanding sector with measures to support
by
"sclerotic"
employment
in the
prescription of gradualism does not act as an
and destruction,
an already depressed creation
for
it
can only reduce destruction
rate.
Introduction
1
1.1
Adjustment Crises
The need
to restructure the economy's productive system
and reallocate
factors
is
common
to a variety of experiences in the developing world, as countries face changing external
opportunities or attempt to liberalize their economic system. Such adjustment episodes are
often times of
crisis.
Those
crises
have important financial and
political
dimensions but,
MTT and NBER and CEPREMAP. An earlier version of this paper was presented at
Seventh Annual Interamerican Seminar on Economics in Mexico City, November 1994. We
are grateful for the comments we received at the conference, and from Olivier Blanchard, Manuel Santos,
Rodrigo Valdes and seminar participants at CEPREMAP (Paris), IDEI (Toulouse), and the Tinbergen
Institute (Amsterdam). We also thank Rodrigo Valdes for excellent research assistance. Ricardo Caballero
thanks the National Science and Alfred P. Sloan foundations for their financial support. Part of this paper
was written while Ricardo Caballero was a Visiting Scholar at the IMF.
'Respectively,
the
NBER
at a
more fundamental
place in an orderly
Adjustment
bears the
full
and
level,
they can be the
efficient
crises are generally characterized
in the
new
of a failure of restructuring to take
manner.
burden of the shock and
and investment
symptom
by an existing productive structure that
faces extensive destruction, while the
structure remains excessively timid. In the
pace of creation
common
experience
of countries subject to adverse external shocks, for example, sharp contractions in non-
tradables are typically only followed slowly by growth in the tradables sector, giving rise
to an acute
employment problem. Brisk trade
lead to such a
crisis,
liberalization,
with an immediate contractionary
effect
it is
often feared,
on importables that
counter-balanced by the gradual growth in exportables. Another example
is
may
is
also
hardly
provided by
the recent experience of reforming economies in Eastern Europe, where the immediate collapse of the old state sector
new
was typically only accompanied by a sluggish emergence of the
private sector ultimately intended to replace
As a consequence
ation,
of the
wedge between sharp immediate destruction and sluggish
an employment problem develops.
Many
tracting sectors find themselves either in overt
much
less attractive activity in
employment
1.2
A
We
focus
crisis
it.
workers
who
cre-
lose their jobs in the con-
unemployment or being forced to take up
the informal sector.
When
the state sector
is
involved, the
often leads to a surge in idle public-sector jobs.
Micro-Structural Approach
on supply-side aspects of adjustment, commonly viewed
the productive structure
as the process
by which
— the sectoral composition of output and allocation of
factor proportions
use — adapts to a new environment.
factors,
the techniques and relative
in
Because of the embodiment of technology in
capital, skills
1
and the organization of work,
restructuring usually necessitates a reallocation of factors of production, which involves the
1
See, e.g., Mussa (1978)
aspects of adjustment.
Our
and Neary (1982)
for seminal contributions to the analysis of the supply-side
general notion of adjustment abstracts away from the specifics of the environment change that gives
it.
We do not address, for example, the exchange rate and stabilization dimensions of external
adjustment in developing countries; the sequencing of reforms in trade liberalization; or privatization and
the behavior of the state sector in the Eastern-European reform experience. One aspect of our approach
that makes it less directly applicable to the Eastern European case is the profit-maximizing behavior we
assume for the contracting sector, which does not capture the nature of decision-making at the level of state
enterprise. For analyses that address this issue, see, e.g., Chadha, Coricelli and Krajnyak (1993), Aghion
and Blanchard (1993) and references therein. See also Shimer (1995) and Atkeson and Kehoe (1993) for,
respectively, search and moral hazard models of reform sharing several of the positive implications of our
model.
rise to
— with
destruction of existing production units that developed under the old environment
the accompanying loss of jobs and scrapping of capital
better adapted to the
We
new
— and the creation of other units
one.
analyze market impediments to the restructuring process that lead to adjustment
A
crises.
very promising approach, we believe,
is
to build our macro-analysis
micro-structure that characterizes exchange in factor markets.
We
on the
focus on the general
incomplete-contracts problem of appropriability in labor and capital market transactions,
which captures the essence of a broad variety of practical obstacles to well-functioning
Appropriability has been identified as a principal dimension of Coasian (1937)
markets.
transaction description in the
New
Institutional
and 1985; Klein, Crawford, and Alchian,
problem
ment
1978; Hart
arises in the joint use of capital
and labor
required. In the absence of complete
is
Economics
and Moore, 1988). In our context, the
is
who may engage
Hammour
appropriability for the cyclical
characteristics of adjustment crises to appropriability problems,
specificity
in opportunistic
behavior of investment, job flows, unemployment, and wages. In this paper,
in times
specific invest-
and binding contracts, investment
sunk (Grout, 1984). In Caballero and
show the general-equilibrium implications of
when
for production,
gives rise to quasi-rents that are appropriable by workers,
behavior once investment
literature (Williamson, 1975
(1994),
we
and long-run
we
trace major
which gain particular bite
— such as adjustment episodes — when high rates of investment and job creation
are required.
It is
to
important to emphasize the prevalence of the appropriability problem, which applies
much more than simply job-specific human capital. There are a number of levels
capital
and labor
political level
interact
— each
put to
level,
its
level imprinting its
At the
may have
is
specific training
level,
which
the aggregate
investment. Investment
its
value
if
At the individual worker-firm
and organizational investment
level of unions, or less formal
in the
worker and
worker coalitions with insider power, assets
to be put to alternative use outside the firm or even the industry to escape the
scope of the union.
exposed.
the union
own form of specificity on
best alternative use, possibly by other users.
the most exposed
his job.
level,
understood as the degree to which the asset would lose of
"specificity," here, is
it is
— the individual worker-firm
at
At the
It is
then the totality of firm- or industry-specific assets that
may be
and labor
interest
highest level of political interaction between capital
groups, even more generic assets become appropriable through such measures as the right
to strike,
employment protection
the possibility
—
legislation,
redistributive corporate taxation, or even
historically quite relevant in the developing
world
— of outright state
expropriation.
2
Unbalanced Restructuring,
1.3
Sclerosis,
and Wage "Rigidity"
Problems of appropriability can have highly disruptive consequences on the functioning
of labor
and capital markets during adjustment. They generally induce the sluggish job
creation and excessive destruction that characterize adjustment crises,
employment problem. The prospect
for firms of
effectively increases their investment costs,
newly expanding
having to yield part of the value they invest
and depresses investment and job creation
if it is
in the
Given the depressed rate of creation, the economic system ought
sectors.
to keep destruction in the contracting sectors at a low
the creation rate,
and the ensuing
enough
to avoid wasting resources. Instead, the
jobs extensively during adjustment. At a time
excessive job destruction
and more
difficult
when
level
commensurate with
economy tends
intense creation
to destroy
and hiring are needed,
employment opportunities are the system's
endogenous way of restraining the bargaining position of workers in their transactions with
firms to avoid
wage surges and guarantee the return on investment required by
markets. Workers
who
in the formal sector
loose their jobs because of the gap between creation
and destruction
employment
in the informal
must
sector, or find themselves
is
financial
either take
orrmuch
less attractive
openly unemployed when some form of unemployment insurance
available.
Paradoxically, despite excessive destruction at the outset of adjustment, appropriability
also induces "sclerosis" in the adjustment process.
Sclerosis characterizes a productive
structure that, ultimately, does not adjust sufficiently to
sectoral composition of output, adopting
new
conditions
new techniques, changing the
— by changing the
capital-output ratio,
In the presence of appropriability, the ultimate extent of restructuring
etc.
is
hampered by
the high effective creation costs brought about by transactional inefficiencies.
The observed wage behavior
differs
implied by the micro-structural inefficiencies
from the fixed wages assumed in traditional models of adjustment
(e.g.,
we
focus on
Lapan, 1976;
Neary, 1982; Edwards, 1988; Edwards and van Winjbergen, 1989). Those models assume
that wages are fixed in terms of the expanding sector's good, or a basket with a large share
of
it
— a necessary condition
adjustment, however,
is
for excessive destruction. Empirical evidence
on wages during
not supportive of that assumption. In her careful description of
See Thomas and Worrall (1994) for an analysis of appropriability through state expropriation in the
case of foreign direct investment.
Latin American labor-market legislation, Cox- Edwards (1993) finds that
minimum- wage
laws have become largely nonbinding in Latin America today.
In comparative studies,
authors such as Fallon and Riveros (1989) or Horton, Kanbur and
Mazumdar
great degree of
wage responsiveness during the adjustment experiences of a broad sample
of developing countries.
In our model, wages
although rigidity
appear
may move with
(1994) find a
will still
exactly the
same
in
may
look quite flexible during adjustment,
more subtle ways. In extreme
flexibility as in
an
efficient
cases,
we show
that wages
economy, but, in order to do
so,
they require the "quantity" movements of high destruction and underemployment usually
associated with rigid wages
— a phenomenon we
market problems may disrupt quantities but leave
show that wages are characterized by a mixture
of
which
is
little
trace
of "covert"
on
prices.
and "overt"
other words, labor
More
generally,
rigidity,
the degree
in labor
variability,
a better measure of the appropriability
markets can be found in the degree of labor market segmentation. Indeed,
when
the degree of asset specificity and appropriable quasi-rents differ across sectors
as
normally the case,
is
we
ultimately determined by the structure of creation costs.
Rather than looking at empirical wage
problem
call covert rigidity. In
for
example, between the formal and informal sectors
—
it
—
is
very natural that wages also differ in equilibrium. Appropriability can thus very naturally
account for labor market segmentation, which
in the developing world (e.g.,
1.4
By
is
a preponderant feature of labor markets
Lopez and Riveros, 1989; Fallon and Riveros, 1989).
Policy Responses
the nature of their driving assumption, fixed-wage models often disassociate the treat-
ment of the destruction margin
dynamics
— driven
by fixed wages and their ad hoc adjustment
— and that of the creation margin — driven by investment dynamics. This typ-
ically results in
destruction,
a dichotomy between a "short-run" analysis of the
and a more
ciation, fixed-wage
different policies
effect of fixed
wages on
neoclassical "long-run" analysis of creation. Because of this disso-
models do not naturally lead to an integrated view of the joint
on the creation and destruction margins. In
fact,
effect of
the popular prescription
of gradualism in the context of those models can be quite sensitive to the fact that its effect
on creation
is
typically not explicitly modeled. In contrast, one of the central features of
policy analysis in this paper
in
is
the emphasis on the fact that policies targeted at distortions
one margin typically have counter-productive implications
margin.
The
real test
is
for distortions in the other
whether gradualism can close the wedge between creation and
destruction to help redress the transitional employment problem.
From
that point of view,
we
an
find little reason in our analysis for
effective "synchronizing" effect of
Gradualism can only reduce the destructive
creation and destruction.
by also reducing the economy's already depressed creation
constitute,
by
itself,
an effective solution
in
rate,
terms of economic
gradualism on
effect of
adjustment
and thus does not seem
to
efficiency.
Policy analysis must go beyond the gradualism versus cold turkey debate, and examine
the managed-adjustment policies needed in face of unbalanced restructuring. Despite the
fact that the
economy's
ills
we
are ultimately institutional in nature,
focus, for a
ber of reasons, on macroeconomic rather than institutional cures. There often
government can do to directly remedy incomplete-contracting
When
social institutions
feasible, or undesirable
of
macroeconomic
and
legislation are concerned,
because of
policies
distortions in the creation
is
its social
On
in turn
the
inefficiencies in transactions.
may be
either politically in-
and distributional consequences. Our analysis
and destruction margins.
ation calls for a creation incentive
may
is little
conducted in terms of canonical policies that directly address
the creation and destruction margins
intensified hiring
change
num-
is
(e.g.,
It is
On
clearest.
there that the interaction between
one hand, the depressed rate of cre-
an investment tax credit or an export subsidy), but
put pressure on the labor market and exacerbate destruction.
the other hand, excessive destruction-calls for job-protection policies
subsidies in the contracting sectors or "gradualism"), but those
by raising the opportunity cost of labor.
What
needed
is
addresses distortions in both margins simultaneously.
in the presence of appropriability requires a
panding sectors and employment protection
debate on gradualism, the implication
is
We
is
may
(e.g.,
employment
discourage creation
an integrated program that
argue that healthy adjustment
combination of creation incentives in the exin the contracting ones.
that a gradualist approach
Going back to the
must be supported by
vigorous creation incentives to be effective.
The
rest of this
paper follows in four sections.
First,
we develop
in section 2 a simple
model of adjustment that captures supply-side aspects of adjustment and of the disruptive
role of appropriability. Section 3 turns to the analysis of the traditional gradualist prescription,
and section 4 derives the optimal
5 presents
2
To
policies that
can restore
efficiency. Finally, section
some concluding remarks.
Appropriability in a Simple
fix ideas,
we
Model
of Adjustment
take the example of trade reform as a paradigmatic case of adjustment.
start with a simple
model that presents our perspective on restructuring
in the
We
adjustment
process and of the disruptive effect of appropriability. In order to have a simple solution, we
assume that adjustment
is
linear in all its aspects. This results in
such as instantaneous reallocation in the
and
linearity,
which we relax
in sections
4.
A
2.1
features,
economy, and degenerate destruction more
However, our main results do not hinge on
generally.
3
efficient
some unpleasant
Simple Model of Adjustment
Our economy has two goods,
X
(
"exportables" ) and
M
(
"importables" )
.
Importables are
used as the numeraire. The representative infinitely-lived household has linear
utility:
f°[Cx {s) + Cm {a)]e<'-^ds.
(1)
Household labor supply
perfectly inelastic, with a labor force equal to L.
is
Both goods are produced with
units in such a
way
similar Leontieff technologies.
that, for each of the goods,
We
choose measurement
one unit of sector-specific capital and a
worker combine in fixed proportions to form a production unit and produce one unit of the
good.
Initial conditions.
tional trade
and
We assume that before liberalization the economy is closed to interna-
in long-run full-employment equilibrium. Since
with respect to
X
initial allocation
of factors
and
M,
is
both goods have pre-liberalization prices equal to
arbitrary.
We let L^
and L~ stand
Liberalization. Let the international relative price of
which
=
0.
is
given for the country.
It
is r,
We assume that
in the
M
An
unanticipated
X
with respect to
M be
6*
>
1,
trade liberalization takes place at
full
the capital account
is
open, and that the world interest rate
the same as the domestic discount rate. Since consumers are indifferent between the
two goods, they specialize
in
good
M.
Producers, on the other hand, would like to specialize
good X. The cost of setting up a new production unit
is Cq,
cost that stands for physical capital as well as organizational
H
employment
amounts to exposing domestic consumers and producers to the international
relative price.
in
for
and the
1
X sectors, where a minus-sign superscript denotes pre-liberalization quantities.
and
t
consumers are indifferent
(t)
takes place in the
X sector, with
all
which
is
a generic investment
and training
costs.
Creation
workers being hired from the unemployment pool
17(f).
Appropriability.
The
appropriability of investment
is
an incomplete-contracts problem
that can
afflict
and the worker
there
may
the efficiency of the job-creating transaction between a firm that invests cq
it
When
hires.
not be a
way
part of the investment
to write a binding
is
"specific" to the
production unit,
and complete contract to ensure that
all
the
quasi-rents attributable to the investment go to the firm.
To capture
this idea,
we introduce a parameter
investment whose quasi-rents are appropriable. To
<t>
6
that measures the share of
(0, 1]
fix ideas,
consider a job that requires a
$1,000 investment, of which $700 must be spent on equipment that can be resold at cost and
$300 on specific training to the worker that
is
operates only at the individual firm-worker
binding contract can be signed.
If
When
of no use elsewhere.
level,
the parameter
<f>
effectively raise
equal to 0.3
the worker belongs to a powerful union, or
protected by legal restrictions, dismissing him and selling the machine
and would
is
appropriability
if
if
his job
may be more
no
is
difficult
<j>.
Incomplete contracting creates a flow of appropriable quasi-rents that must be shared
by the two parties through a bargaining process. Let
a firm and a worker
who meet
at time
to the continuous-time generalized
the worker
The
and
(1
— (5)
surplus IT(£)
and the firm can
is
solution.
A share
/3
(0, 1)
of the surplus goes to
equal to the value that the match creates above
effectively claim as their best alternative.
(1
— <P)cq
if it
drops out of the match.
worker's best alternative in the calculation of the match surplus
unemployed and searching
for
another job.
[9*
We
what the worker
Because of the appropriability
of its investment
w(t), equal to the opportunity cost to a worker of remaining
It is
G
goes to the firm.-
problem, the firm can only recover
The
denote the surplus over which
must bargain. Bargaining takes place according
t
Nash
IT(t)
is
the shadow wage
on the job rather than turning
can thus write the match surplus
- w(s)] e'^^ds -
(1
-
(j>)co,
t
>
as:
0.
equal to the present value of the future revenue 6* from production minus the worker's
shadow wage
tu,
after subtracting the protected part of the creation cost cq.
the opportunity cost w(s) for the worker of holding the job
probability H(t)/U(t) of finding another
is
match times the part
In turn,
equal to the instantaneous
@Tl(t) of the
match surplus
he would get there:
(3)
Equilibrium Conditions.
We
w(t)
= ^f3U(t).
assume
free entry for
8
new production
units.
This means
,
that, as long as entry
entrant's share of the
taking place, the cost of creating a production unit
is
is
equal to the
match surplus plus the part of investment that can be protected. 3
Rearranging this condition, we get
4>co
(4)
The
it is
match surplus
cf>
=
0,
is
called
(l-P)U(t).
match surplus must compensate
firm's share of the
investments
=
on to make. This shows
created
if
and only
if
there
is
the free-entry condition becomes H(t)
=
exactly for the unprotected
it
clearly that,
under
free entry,
a positive
an appropriability problem. Otherwise,
0,
which by
(2) is
if
the standard condition
that equates Cq to the present value of quasi-rents.
Solving out for
II(£),
equations (3)- (4) imply the following equilibrium conditions for
all
t>0:
[0*
(6)
- w(s)] e^^ds,
L m (t) = L"
if
w(t)
<
0<L m (t)<L-,
if
w(t)
= l;
Lm (t) =
if
w{t)
>
U(t)
(7)
1,
= L- Lx (t) - L m (t),
H
(8)
0,
1;
(t)
= dL x {t)/dt,
where
5(()
(9)
(l
+ i)o,
W §§^,
.-^
=
(10)
and
(ii)
The
=
first
equation governs free entry in the economy, where the "effective" creation cost
c(t) is distorted
by a factor
6.
4
This distortion
is
due to the problem of appropriability. In
equilibrium, the part of investment whose quasi-rents are appropriated by workers
The "appropriation" parameter
b given
by
(11)
is
is bco-
increasing in the degree of appropriability
3
FonnaUy, the free-entry condition is cq = (1 — /3)II(t) + (1 — <f>)coFor creation to be economically sensible, we must assume that the effective creation cost
so as to be covered at least by the present value of revenues
i.e. (1 + 6)co < 0' /r.
—
is
low enough
<p
and
bargaining power of workers
in the
It states
/?.
The second equation
M sector
that production units in the
is
the exit condition.
are scrapped or preserved depending
on
whether their associated quasi-rents are negative or positive, which in turn depends on
whether the shadow wage w(t)
The
latter
is
is
greater or smaller than 1 (productivity in that sector).
given by equation (10), which follows from (3) and
(4).
The
third
and fourth
equations are accounting equations that define unemployment and hiring.
An attractive feature of the decentralized-bargaining equilibrium in
it
this
economy
is
that
converges to the efficient competitive outcome as the appropriation parameter b goes to
In the limit, the effective creation cost in (9) becomes undistorted,
zero.
wage
down
in (10)
becomes whatever
is
and the shadow
needed to clear the labor market and bring unemployment
5
to zero.
The Nature of Unemployment. The free-entry condition
respect to
t
and solved
for the constant
on investment. Expression
can be differentiated with
(5)
shadow wage that guarantees the required return
(10) can then be used to calculate, for a given level of hiring,
the level of unemployment needed to yield the required shadow wage in equilibrium.
We
get
(12)
w(t)=e*-r(l + b)co;
13 >
«W = *--K?+i) co g ">-
<
It is clear that, in
new
locations,
an economy with important needs to restructure and hire workers in
unemployment
appropriability. If
is
the result of the positive match surpluses that result from
unemployment were
zero, workers
would find
it
infinitely
the positive surplus of an alternative job. This outside alternative would
wage
infinite (equation 10),
easy to capture
make
their
shadow
which renders job creation prohibitively expensive and generates
unemployment. Unemployment
is,
thus,
an equilibrium response of the economic system
that restrains the bargaining position of workers in the presence of appropriability and
preserves the profitability of investment.
and intense gross hiring require high
wages and keep them at a
As equation
transitional
level that
(13)
makes
clear,
times of adjustment
unemployment to prevent surges
makes job creation pay
in
shadow
off.
Equation (10) turns into the full-employment equation once we multiply both sides by U(t) and set
b
= 0.
10
and Unbalanced Restructuring
Sclerosis
2.2
Sclerosis.
We can now solve
wage together with the
in
which sector
The
M
inaction range
enough
1 for
for the
economy. Solution (12)
for the evolution of this
exit condition (6) defines
an inaction range
in
for the
shadow
parameter space,
remains profitable and no adjustment takes place following reform.
is
<
given by 0*
+ r(l + b)co,
1
which clearly indicates that
it is
not
world relative price 6* of exportables versus importables to be greater than
restructuring to occur. 6* must be sufficiently larger than
1 to justify
spending the
additional creation costs.
Note that there
the inefficient
is
a segment 6* G
economy does not
(1
+ rco, 1 + r(l + b)co]
adjust, while efficiency
greater the appropriation parameter
b,
the larger
is
this
of the inaction range where
would require adjustment. The
segment of the inaction range.
Thus, appropriability can prevent adjustment from taking place altogether in cases when
efficient to adjust.
it is
we term
This
is
the
first
manifestation of an effect of appropriability that
sclerosis.
Sclerosis arises in the
ficiently in
response to
economy whenever the productive structure does not adjust
new
conditions
— by changing the sectoral composition of output,
adopting new techniques, changing the capital-labor
ratio, etc.
In this particular case, ap-
propriability can cause the productive structure to be locked into its past
response to the opening of trade, despite the efficiency of doing so.
is
suf-
and not adjust
The reason
in
for sclerosis
the high effective creation costs induced by appropriability (equation 9), which reduce
the incentive for creation as well as the hiring pressure on wages and destruction (equation
12).
The
fact that sclerosis takes the
the degenerate
incomplete
extreme form of
initial distribution in
— rather than
fully
M
the
aborted
sector.
fully
2.4, or
is
due to
Generally speaking, sclerosis leads to
— restructuring, as
technology model analyzed in sub-section
aborting restructuring
the case in the heterogeneous-
is
when the M-sector
exhibits diminishing
returns to labor.
Unbalanced Restructuring. Outside the inaction range, the
and dynamics are governed by equations
(14)
dU(t)/dt
(7)
and
= -H(t)
(8)
with
11
with
M sector
Lm (t) =
J7(0)
= L".
0:
is
instantly scrapped
Taking (13) into account, the solution to
U(t)
(15)
this
equation
= L^e
*»
is,
for
t
>
0:
«,
Lx (t) = L-U(t),
Lm {t) =
The whole
M sector
0.
scrapped instantly following reform, leading to a jump in unemploy-
is
ment. However, creation in the
X sector
is
progressively over time. Asymptotically, the
gradual and can only absorb the unemployed
economy
adjusts fully
and reaches
employ-
full
ment.
This path has very
outcome corresponds
the efficient
economy
inefficient characteristics.
to the limit
is
when
6
As we have
goes to zero.
seen, the efficient competitive
easy to see that adjustment in
It is
instantaneous and complete at time
t
=
This means that creation
subject to appropriability problems exhibits sluggish creation.
and destruction are out of balance and lead to
instantaneous efficient adjustment
show
in the
is
transitional
unemployment
on creation and destruction does not hinge on
this.
its efficient level is
Although
we
effect of appropriability
Moreover, we show that, at the outset
of reform, appropriability generally leads to insufficient creation
efficient
costs.
not a very attractive feature of this simple model,
convex model of section 3 that the de-synchronizing
compared the
economy
In contrast, an
0.
and excessive destruction
outcome. The fact that, in the present case, destruction
is
equal to
only due to the maximal, instantaneous rate at which the latter takes
place.
2.3
Covert
Wage
Rigidity
Although this model exhibits the transitional unemployment usually associated with
wage models
may appear
(e.g.,
Lapan, 1976; Neary, 1982; Edwards and van Winjbergen, 1989), wages
quite flexible. In fact, in the special linear case
the same as what they would be in an efficient economy.
actual
fixed-
wage paid by exporting
we
To
are analyzing, wages are
see this, denote
firms (the only firms after liberalization).
not equal to the opportunity cost of labor, w, for
we need
to
add the part of the match
surplus that goes to the worker, which in flow terms amounts to rbcQi
wx = w + rbco = 6* — rcQ.
12
Note
wx the
that wx is
by
This expression
efficient
is
independent of
economy where
b
=
adjustments
ployment
—
in the
and
therefore exactly equal to the
is
it
from surging beyond that
a hidden form of rigidity that manifests
As we
In Caballero and
will see in the
Hammour
itself entirely in
(1994),
rigidity,
is
inefficient are the
wages
level
quan-
during times of intensified
we dubbed
3,
the equilibrium quantities that
this
phenomenon
observed wages exhibit, more generally, a
restrain
them from being even
consistent with the empirical evidence
is
covert rigidity.
which partly keeps them above their
and partly requires quantity adjustments to
variability of
an
wage appears adequate, but actually harbors
convex model of section
mixture of "overt" and "covert"
The
in
form of unbalanced restructuring and high transitional unem-
— needed to restrain
it.
what
as in the efficient outcome,
hiring (see equation 10). In other words, the
support
wage paid
0.
Although the wage behaves
tity
ft,
efficient level
higher. 6
on wages, which
generally exhibit a great degree of responsiveness in case studies of adjustment (Fallon and
Riveros, 1989; Horton, Kanbur, and
Mazumdar,
1994).
However,
caveat for the interpretations of those findings, from which
it
raises
an important
some authors have concluded
that adjustment crises are not due to malfunctioning labor markets. In their recent
Bank study
World
of the role of labor markets in the process of adjustment, for example, Horton,
Kanbur and Mazumdar (1994) conclude:
[Tjhere are three possible explanations as to
during stabilization. The
cause of real wage
not working well be-
The evidence presented by the
case studies certainly
that the labor market
does not favor the view that real wages were
employment. Even
for Chile,
the longest, real wages
nations: aggregate
fell
persist
is
first is
rigidity.
why unemployment may
and therefore led to un-
where unemployment was highest and persisted
dramatically.
[...]
demand feedback from
output market imperfections.
rigid,
[T]his leaves the other
two expla-
declining real wages to output
and
(Vol. 1, pp. 19-20)
In the linear model, the shadow wage rises in terras of importables following reform. This is very
and is precisely the reason why the
sector contracts. In our convex model, this shadow wage
remains unchanged while the
sector is being depleted, and jumps to its long-run value when the process
is completed.
For the
sector, the product-wage in both models also generally rises in the long run. This effect is
only due to the fact that creation costs were assumed fixed in terms of importables, thus making it cheaper
natural,
M
M
X
Had we fixed those costs in terms of exportables instead, the long-run product
sector would have been unchanged. In the transition, however, the product-wage in the
immediately to its long-run value in the linear model, but remains temporarily lower in the
to create following reform.
wage
in the
X
jumps
convex model to induce
sector
X
intense creation.
13
Our model
clearly
shows that looking at the path of
the extent of labor market rigidity
exogenously fixed
wages
real
in isolation to assess
— which comes from the practice of modeling wages as
— can be quite misleading.
Variable real wages do not necessarily imply
that the labor market functions adequately. Other forms of empirical evidence
is
needed
to assess the workings of the labor market. Looking at the degree of segmentation in labor
markets
— which
1989; Fallon
problems.
preponderant
and Riveros, 1989)
segmentation
2.5,
is
is
Another
in the developing
world
— may constitute a better
(see, e.g.,
Lopez and Riveros,
As we explain
test.
a very natural feature of a labor market
afflicted
in sub-section
by appropriability
can be gleaned from the institutional features of
class of evidence
labor markets that characterize the more "visible" aspects of appropriability. Cox- Edwards
(1993) gives a careful account of the predominant institutional distortions in Latin
American
labor markets. These include employment protection laws, payroll taxes, and antagonistic
labor-management relations that encourage confrontation and costly settlement procedures.
Consistently with the authors cited above, she concludes that
the
list
of the
most pressing labor market
issues in Latin
minimum wages
are not on
America today.
Productivity Cleansing During Adjustment
2.4
Our
structural approach, based on the
the adjustment process further, once
it is
embodiment of technology, allows us to characterize
we
allow for heterogeneous technologies. In that case,
simple to see that the shock of liberalization affects the least
the most. Thus,
if
gives rise to a
phenomenon of productive
which finds growing support in the empirical
To develop
this point,
we modify
tion of production units in the
efficient
productive units
"cleansing" in the economy,
literature.
the above model and assume that the initial distribu-
X and M sectors
is
heterogeneous in the technologies in use.
For each production unit, we use a as an index for the outdatedness of the technique in use
and assume a productivity oil— a. Technology
to a
available today in
both sectors corresponds
= 0.
The distribution of production units
across techniques
is
given by the history of technol-
ogy adoption (see Caballero and Hammour, 1994) and government intervention. For each
sector,
we assume a non-degenerate
the scrapping margin
a~
is
initial distribution
over the range a
determined by the zero profit condition
1
— a~ =
14
it;
,
€
[0,a~],
where
and the shadow wage
w~ =
is
1
— r(l + b)cQ
determined in the same way as in equation
(12).
7
Consider what happens after liberalization. The interesting phenomena here arise inside
the old inaction range (6*
there
is still
<
1
+r(l +
M
a complete and instantaneous scrapping of the
followed by instant scrapping of
condition, which
price.
where we assume the economy
b)co),
is
now
all
otherwise
Liberalization
production units that do not satisfy the
different for the
The new scrapping margins
sector.
is,
new
is
zero-profit
two sectors because of the change in their relative
M and X, respectively, are given by
for sectors
l-a m = w
(16)
and
6*(l-ax )
(17)
where
w
is still
= w,
given by the free-entry condition in the
after the initial scrapping are as before.
X sector (equation
The unemployed who
lost their
shock of liberalization are hired gradually into the expanding
12).
Dynamics
job after the
X sector,
initial
while production
units in both sectors that were not scrapped initially subsist in the long run.
Following liberalization, the scrapping margins (16)-(17) tighten in both sectors by the
amount of increase
in the product wage. This
phenomenon of
"cleansing" of the productive
structure following reform, that results in an improvement in average productivity, finds empirical
support in the recent literature
plants during the 1979-86 period). It
margin tightens more markedly
(see, e.g., Liu's
is
for the
also
1993 study of Chilean manufacturing
immediate from (16)-(17) that the scrapping
M sector than
implies that productivity cleansing following reform
is
for the
X sector (a„,
likely to
< ax ). 8
This
be more extensive in im-
portables than in exportables, which also finds empirical support (see,
e.g.,
Tybout, De
Melo and Corbo, 1991; Haddad, 1993; Corbo and Sanchez, 1992; Marshall, 1992).
With heterogeneous
The shadow wage
technologies, the problem of sclerosis discussed in sub-section 2.2
determined by the free-entry condition, even though we assume that the economy
a worker is still
to find a potential entrant, even if such a threat is not realized in equilibrium.
In fact, the only reason the scrapping margin tightens and the product wage rises in the
sector is
because we have assumed that creation costs are fixed in terms of importables. Had we assumed them fixed
in terms of exportables, the product wage in the
sector would have remained unchanged. Had we also
assumed convex adjustment costs, the X-sector wage would have fallen in the transition.
at time zero
is
is
in long-run equilibrium with zero entry. This is because the threat point of
X
X
15
,
gains a technological dimension.
One can
and
see from equations (12)
(16) that a greater
degree of appropriation b lowers the shadow wage and loosens the scrapping margin a m in
M
the
sector. All production units with a
€
—
[tcq
(6*
—
l),r(l
+
b)co
—
(0*
—
despite the fact that they ought to be scrapped in an efficient world. Sclerosis
just a
problem of producing the wrong goods, but
also of
producing
all
subsist
1)]
is
goods with
now not
inefficient
technology.
Unemployment, the Informal Sector and Segmented Labor Markets
2.5
In developing countries, where
unemployment
benefits are sparse at best,
unemployment
is
typically "hidden" in the form of participation in low-productivity informal-sector activities.
Instead of a surge in open unemployment, the transitional employment problem
may
take
the form of a surge in informal sector employment.
The segmentation
of labor markets between the formal
naturally in our context,
if
and informal sector
arises very
informal-sector technology requires less asset specificity or allows
workers to avoid entering into a bilateral transaction with capital owners. 9 The simplest
case in which this happens
and assume
is
when the technology requires no
that, in the informal sector, "workers can
that requires no capital and yields a
The only
all
We modify our model
become self-employed
— possibly quite low — productivity
as close as possible to the above model,
and assume that
capital.
we denote
total informal sector
7r
in
<
an activity
1.
To
stay
employment by U,
workers hired in the formal sector come from the informal one.
modification to equilibrium conditions (5)-(ll) comes from the fact that
must add the productivity of informal-sector
we
activity to the opportunity cost of labor in
(10):
w_,^ =
,
Hit)
W)
bC0+lr
-
This means that the level of informal sector employment U(t) needed for a given
hiring
is
now
level of
higher than in equation (13), and equal to
PW" y-.)5(i+»)» *ft>The
possibility of engaging in informal-sector activity strengthens workers' threat point,
which must therefore be
offset
by an even higher
level of
U
to maintain firm profitability.
For an analysis of some of the consequences of labor market segmentation in small open economies, see,
e.g.,
Agenor and Aizenman (1994).
16
The dynamic equation
(15) characterizing U(t)
becomes
U(t)=L^e-
A
more productive informal
The
-o
down
sector further slows
divide between the formal and informal sector
is
10
the adjustment process.
only one aspect of the segmentation
that can arise in the presence of appropriable investment. Naturally, as
any sectoral differences in the degree of asset
specificity
can cause segmentation. Moreover,
many differences in the determinants of profitability that
can interact with appropriability and lead to wage
of the
we mentioned above,
arise after some
differentials
model with heterogeneous technologies developed
— as
is
investment
is
sunk
the case, for example,
in the previous sub-section.
11
Gradualism
3
Our approach
to adjustment crises provides us with a framework for the study of alternative
policy responses. In this section
and
in the next
we turn
we concentrate on the popular
prespcription of gradualism,
to the question of optimal policy design.
Our
analysis of gradualism
provides an opportunity to generalize to a convex model of adjustment the central results
we obtained
We
section.
have seen that appropriability induces two distortions in the transition: depressed
creation
crisis.
model of the previous
for the linear
and excessive destruction. The wedge between the two
Much
transitional
gives rise to an
of the popularity of gradualism derives from the idea that
employment
crisis
assumption that creation
slowing
down
is little
affected. In fact, there is little reason
both creation and destruction
adjustment in the process
can
relieve the
by slowing down destruction, presumably under the implicit
the transition will not also exacerbate the sluggishness of creation.
effect of
it
employment
— are ambiguous, and
why
slowing
The
net efficiency
down
— and of delaying the net benefits of
likely to
depend on the
specifics of the
situation. 12
Note that exactly the same
would
from the introduction of unemployment benefits.
we need to make additional assumptions about the out-ofequilibrium uses of the non-appropriable component of capital, including the relation between those uses
and current productivity and location. Most reasonable scenarios, however, lead to wage differentials that
To
calculate the actual
effect
wage
arise
differentials,
are related to the relative profitability of the production units being compared.
There
a class of models characterized by over-investment, where the pace of both creation and destrucand where the welfare-improving properties of gradualism are therefore unambiguous.
A good example can be found in Gavin (1993), who advocates gradualism based on a congestion model of
search. His assumption, that the matching function is a concave function of unemployment only, implies
is
tion are excessive,
17
l
To uncover
the tension between the effects of gradualism on the creation and destruction
The
margins, we analyze two extreme cases.
first
extreme highlights the distortion
creation margin and corresponds to the simple linear model presented above.
destruction in that model
is
down
creation further,
other extreme
is
is
we analyze
Since gradualism can only
sluggish creation.
its efficiency effect is
what would be required instead
The
unambiguously negative. In that case,
"accelerationism."
highlights the distortion
on the destruction margin.
same simple model analyzed above, but now with a
consist of the
Convex creation
creation costs.
the special assumptions
special
that the dominant distortion
destruction side. In that case, gradualism
is
and destruction.
Its
main
effect will
always on the
when the
multaneously slow down destruction and accelerate creation. Because
is
and
always beneficial.
reveal the limitations of gradualism
margins simultaneously, gradualism
is
It
form of convex
costs lead to non-degenerate destruction dynamics,
we make imply
Those two extreme cases
Because
instantaneous in both the efficient and inefficient case, the
only problem with transitional dynamics
slow
in the
it
goal
slows
is
to
si-
down both
unlikely to be an effective "synchronizer" of creation
be to spread out transitional underemployment costs
them
a more general setting, the economy
over time, but
is
will present a
mixture of the two extremes we analyze, with phases of adjustment where
gradualism
beneficial,
is
gradualism and
distortions
is
unlikely to reduce
its
effectively. In
and phases where accelerationism
is.
This alternation between
opposite, depending on which of the creation- or destruction-margin
dominant, shows more generally the limitation of using a single policy
in-
strument to address distortions on the two margins. Policy analysis should go beyond the
gradualism versus cold-turkey dichotomy, to the design of managed adjustment programs
with multiple policy instruments. This we proceed to examine in the next section.
Throughout
this section,
we model a
gradualist reform as the temporary introduction
of a path of tariffs {r(t)}t>o following reform to protect the importables sector.
the
tariff is
We assume
small enough so the domestic relative price of exportables 6{t) remains above
one:
* (t »
" (TTRJj)
a
t£0
'
In order to reduce technical complications in the analysis,
capital account
is
closed
and that the
interest rate r
is
-
we assume from here on
given by the subjective discount rate
that high unemployment, by facilitating matching, pushes creation above
18
that the
its socially efficient level.
in the utility function (l).
constraint
is
not binding
We
13
(i.e.,
assume parameters are such that the balance-of-payments
desired domestic investment does not exceed desired domestic
saving).
Gradualism with Sluggish Creation
3.1
In this sub-section,
to gradual opening-up in one extreme case that corresponds to
we turn
the simple linear model of sub-section
same when a
the
tariff is
2.1. It is
introduced at time
t
easy to see
=
equations (5)-(14) remain
except that the world relative price 6*
0,
must now be replaced by the distorted domestic price
is
how
the shadow wage
6{i). In particular,
now given by
w(t)
(18)
= e(t)-r(l + b)co.
Assuming model parameters are such that even with gradualism the economy does not
inside the inaction range
fall
unemployment
is
now
>
that 6(t)
C/(t)
= L"e
define two paths of reform.
Jo
The
oco
given by 6
path
is
=
(t)
+ r(l + b)co,
for all
9* for all
t
>
0.
d
\
>
t
It is clear
The
the path of
is
characterized by an
c
cold-turkey path {d (t)}t^o
simply
more sluggish creation
and a smaller
X sector.
destruction
the same in both scenarios. Gradualism in this case has one
Since
1S
from equations (18)-(19) that the gradualist
characterized by lower shadow wages, higher unemployment,
is
0),
£>o.
gradualist path {0 g (t)}%LQ
increasing 6 9 (t) that reaches 6* asymptotically.
c
1
given by
(19)
We
(i.e.
we have assumed
that
we remain
outside the inaction range,
reduce the relative-price incentives to create in the exportables sector.
It
effect,
which
is
to
leads to reduced
creation and, given complete destruction in importables, higher unemployment.
We analyze
define
it
welfare in this
economy from the point of view of allocational
efficiency,
and
as the present value of aggregate domestic income net of investment expenses:
(20)
W(t)
=
J™
[6*Lx (s)
+ Lm (s) - coH(s)) e^^ds.
The difficulty comes from the change in the relative price of importables delivered domestically and
importables delivered overseas that results from a gradual change in the tariff rate. This means that a world
interest rate that is constant in terms of the
good delivered overseas, will not be so in terms of that good
M
delivered domestically.
To avoid changes
in the interest rate,
interest rate to the subjective discount rate.
19
we
close the capital account
and equate the
Although,
in principle,
separable,
we do not
with linear
utility the allocational
much comfort from
derive
ments' ability to redistribute income
in
ever, that, once reform has been decided,
grounds
and
given the severe limitation of govern-
this,
a non-distortionary way.
It
does seem to us, how-
an adjustment path that ranks high on allocational
have to avoid wasted resources
will
and distributional problems are
form of open or hidden unemployment,
in the
rank high on distributional and poverty-alleviation grounds
will therefore potentially
as well.
The appendix calculates
ualist
adjustment
(W9 ),
welfare for three scenarios: cold- turkey adjustment
productive structure intact). First
appropriability problems,
private incentives,
it
which
is
we show
which corresponds to keeping the old
,
that, notwithstanding the disruptions
cold-turkey adjustment
if
will also
W
(Wn
and no adjustment
c
is
due to
desirable from the point of view of
be desirable from a social point of view:
- W"(0) = -
(0)
{Wc ), grad-
[(*•
-
1)
r
clearly positive outside the inaction range.
- r(l + b)co] L-,
14
Turning to gradual adjustment, we show that
W
9
-
(0)
W
c
(0)
=-
f°° W*
-
9 (s))
U9 {s)e- r ^-
t)
ds,
Jo
where
U9 (t)
expression
designates unemployment along the gradualist path.
is
negative
when 6 9 {i)
stays below 6*,
It is clear
and that gradualism
is
that this
inferior to cold-
turkey adjustment. In this extreme linear case, gradualist adjustment can only affect the
creation margin. It slows
unemployment
3.2
down an
already depressed creation rate, and results in increased
as well as a delay in the net benefits of adjustment.
Gradualism with Excessive Destruction
The reason why gradualism
destruction
precisely
is
fails
to have any beneficial effect in the linear
degenerate and happens instantaneously.
on the idea
that,
The
model
is
that
gradualist argument hinges
by slowing down the destruction of the old structure, a gradu-
14
The reason for this is that, outside the inaction range, the private distortion 6co to creation costs
exceeds the social cost of the additional unemployment that results from creating an additional unit (equal
to U/H = bco/w by equation 13, which is less than bco outside the inaction range), while the private value
of a new unit created in the
sector is less than its social value (because the private shadow wage exceeds
the social shadow wage of zero in the presence of unemployment). Thus, a private decision to create is both
X
cheaper and yields greater value from a social perspective.
20
alist
approach can give more time
for creation to build
up the new one, and thus reduce
transitional under-employment.
To capture
this effect,
instead that fast creation
is
we drop the assumption
The shadow wage no
costly.
following reform, precipitating
all
of linear creation costs
longer
jumps
to
its
and assume
long-run value
destruction at that point, but converges over time to
long-term value and brings about a progressive destruction of the
M
its
sector.
In order to highlight the beneficial effects of gradualism on destruction in as simple a
setting as possible,
we make
special assumptions
— on the form of creation costs and on
— that greatly simplify the analysis and imply
the initial distribution of production units
that the destruction-margin distortion always dominates. In this extreme case, gradualism
is
always a beneficial
sub-section,
we
—
albeit far
from optimal
— policy
response.
At the end of
this
more general assumptions.
discuss the implications of
The model. There are many reasons why an economy cannot instantly create a new
productive structure: time to build, slow learning and technology adoption, limited savings,
uncertainty, etc.
To capture some
of those effects,
and assume that the unit creation cost
is
we modify the model
increasing with the rate of creation H{t). This
can be due to standard convex capital installation and learning
to a non-degenerate distribution of potential entrants (see
the reasons explained above,
form ciH(t),
c\
>
we choose
in sub-section 2.1
for the
costs.
Diamond
It
can also be due
1994, chapter
1).
For
marginal creation cost the special functional
0.
Equilibrium conditions (5)-(ll) must be modified. 15 The change in the specification of
creation costs
means that the shadow
price signals (9)-(10) are
c(i)
(21)
now
given by
= (l + b) Cl H(t)
and
w(t)
(22)
Let us
first
=
§^bCl H(t).
characterize adjustment in this
economy without government
intervention.
We show by construction that, as long as 6* is not too large, the transition ends in finite time
T — creation, destruction and unemployment being positive before T, and zero afterwards.
Ongoing destruction requires that wages
M. By
exit condition (6), the
shadow wage
Naturally, as in sub-section 3.1, 6'
rise to exactly eliminate all quasi-rents in sector
will
have to be
1 for t
must be replaced by the distorted
21
<
T, and 6* afterwards
price 6(t) in those equations.
to prevent further creation:
16
{
'
We
[
use
(5), (21)
and
0<t<T;
~_
6*,
t>T.
(23) to solve for the creation rate
l-e- r T -*)], 0<i<T;
e'-i
(
r(l+6)ci
H(t)
(24)
1,
(
=
w(t)
(23)
X
t>T.
11
0,
This can be used to integrate the dynamic equation
using the boundary condition
(8)
Lx (0) = L-:
L x (t) =
(25)
L *" +
\
^fe
t
-e_I
^
(e
rt_
0<t<T;
l)
t>T.
The unemployment
rate along this path can be obtained from (23)
and
(22), taking (24)
into account:
,
s
U(t)
26
=
f
w
bclH(t) 2
< t < T;
~_
,
{
t
| 0,
>
T,
with the destruction margin adjusting to generate the required unemployment.
check that, provided that the technical condition (6*
employment L m
= L — Lx — U
initially
jumps down by
—
1)
£7(0)
<
^(1
+
and then
One can
£) holds, M-sector
falls
monotonically
to reach zero at time T. 17
Finally, the length
ment that the
T of the transition
is
transition be complete at T,
determined from equation (25) by the requirei.e.
LX {T) =
L. It
is
implicitly
determined by
the following equation:
r-i^-3E}±a&(E-«.
(27)
16
Note that, unlike the
1 to 1/6').
The same
linear model, the
shadow wage
holds for the actual product wage
initially falls in
wx /6'
in the
terms of exportables
(it
drops from
X sector, which must fall below 1
to
encourage intense job creation. In fact one can show that, for t < T, w z /0* = 1 — A b)e { Thus, although
the product wage falls in the transition, it falls by less
and exhibits more "overt rigidity
the greater
the appropriability parameter 6 is.
—
,
.
—
We need to check that dL m /dt < for t < T. Using (8), (24) and (26) to differentiate Lm =X-L I -U,
wegetdLm /dt = -H [l - ^(0* - l)e- r(5: -'>l Our condition is satisfied if and only if (0* - 1) < j(l+£).
17
.
22
This model with progressive destruction adds realism and insight, and allows us to
generalize to a convex setting our previous conclusions concerning unbalanced restructuring
and
sclerosis in the presence of appropriability.
and the disruptive
lines
effect of appropriability
The path
problems are illustrated
—
the solid lines correspond to an efficient economy (the limit as b
characterize adjustment in an efficient
Because
More
fast creation is costly, creation
and destruction
results:
(i)
T>T
6
Turning to an
.
inefficient
dT/db >
creation in the
in equation (27); (ii)
in that
reform.
To
economy with
>
b
while
1),
in equations (23)-(27).
economy are
0,
progressive.
unemployment
in the
we obtain the following
X sector
is
measured by T, which can be seen by
L x (t) <
L%(t) for
€
all t
(0,T^],
differen-
which means that
slower that efficient, because of the high effective creation costs
induced by appropriability; 19 (Hi) Sector
=
=
=
(b
18
(where the superscript "e" refers to the efficient-economy benchmark),
transition takes longer than efficient, as
tiating
at t
set 6
0).
importantly, they go hand in hand and therefore generate no
process (equation 26)
i.e.
economy we
The dashed
in figure 1.
correspond to an economy that suffers from appropriability problems
To
economy
of adjustment in this
M
suffers excessive destruction at the outset of
see this, note that, unlike the efficient economy, equations (24)
show that a segment of the
M sector
is
and
(26) taken
destroyed instantly after reform to generate
the unemployment required to prevent wage surges in response to increased hiring. This
is
more progressive destruction that depletes that
followed by a
sector over time; (iv)
wedge created between slow creation and excessive destruction gives
in
unemployment during the
a wasteful surge
transition (equation 26).
new
This model with non-degenerate destruction dynamics adds a
Gradualism.
mension to the welfare-analysis of gradualism.
exhibits sluggish creation
creation even further,
18
rise to
To generate the
figure
and excessive
We
have seen that the
we used the
inefficient
following parameter values:
r
=
di-
economy
destruction. Although gradualism will slow
may help by reducing destruction. Could
it
The
down
the latter effect overcome
0.06, 0*
=
1.2, ci
=
1.5,
and
L~ = L~ = 0.5.
19
To prove
this
we
must have H(0)
<
show that H(Q) < Hc (0). Differentiate free-entry condition (5) with respect to
and T >
into account: dH/dt = - \{'^ + rH, t
\ By contradiction we
first
time, taking (21), (23)
T
H e (0),
smaller (negative) slope
otherwise
<T
we would have H(t) > H'(t)
over the interval [0,7*] because of the
dH/dt than
in the efficient case. But H(t) cannot exceed H'(t) over the whole
e
would contradict the fact that
Since H(t) is decreasing, the fact that H(0) < H°(0) together with H(T) > H°(T°) =
implies that
=
the inefficient path {H(t)} crosses the efficient path {H'(t)} from below at a point t* € (0,T ). This
proves that L x (t) < L%(t) for all t* € (0,5*], for if L x were to exceed L% at some point, that point would
have_to be after t* and would mean that L x (t) > L%(t) from then on. But that contradicts the fact that
Lx(T') < L%(T°) = L.
T>T
interval [O.T*], because that
23
.
:
-i
o
n
in
t in
m
m
tii
ii
x)
i
i
/
/
CM
E
'-P
I
I
I
I
09"0
02'0
OE'G
Ofr'O
OO'O
Ol/O
S'O
fr'O
2"0
-
I'O
q.uawAo|diuaun
ui q.uatiiAo|diU8
[Al
E'O
IT)
a
r-
II
II
_
-i
IT)
-
XJ XJ
.
i:
l
\
\
-
ro
\
\
V
\
^
x.
X.
^
CM
x
N
^^^ v
m
E
4J
>^
X.
"V
^
^w^"S^%
^^^.
-v
^^^^^^
1
001
1
—
.1
1
OB'O
X
1
08'0
1
1
OZ/0
1
1
09'0
uj q.uaiuAo|duja
1
OS'O
E£'Q
frS'O
91'0
uoiq.eajo
80"0
00"0
making gradualism a
the former,
To address
this question,
r
l|r(
The
)
—
I
we
a* j —
write aggregate welfare in this
roo
= tstll±s. +
economy
r
Jo
as
i
[(*'
-
term corresponds to welfare
first
remedy?
useful
l)(Lx (t)
if
- L-) -
U(t)
- - Cl H(ty e~ rt dt.
no adjustment takes place. The second term
is
equal to the improvement in welfare due to adjustment: the increase in the output of units
that are reallocated from the
transitional
The
unemployment, minus
loss of
That
creation.
M to the X
minus the
loss of
M-sector output due to
total creation costs.
output due to unemployment can be considered an "unemployment-cost" of
is
because, by (26), lost output per unit-flow of creation
(28)
H(i)=
Note that
sector,
this average unit-cost
costs introduced
t<T
bClH(t)
>
is
given by
-
exactly equal to the distortion to marginal creation
is
by appropriability. In other words, the appropriability distortion gives an
adequate private signal of the unemployment-cost of creation that appears at the social
level,
side
up
is
to one caveat: the left-hand side" of (28)
is
an average cost, whereas the right-hand
a marginal cost. Because of that subtle difference, private signals do not
full social
reflect the
costs of creation.
To explore
the advisability of a gradualist path of {9(t)},
that maximizes aggregate welfare.
The
we look
first-order condition to the
for the
problem of finding the
optimal path {H°(t)} of creation as driven by the optimal policy {Q°(t)}
c(t)
Compared with the
+ bciH(t) =
f
[6*-
1]
e-^'-^ds,
t
path {9°(t)}
is
< T°.
decentralized free-entry condition (5), the social effective marginal cre-
ation cost exceeds the private cost by a term bc\H{t) which precisely reflects the difference
between marginal and average costs discussed above.
The
fact that the social effective cost of creation
that slowing
down
creation
is
o
beneficial.
e (t)=
i
is
higher than the private one, implies
One can show
that the optimal gradualist path
«*-rk(f-l). 0<t<T°;
1
'
t>T,
e\
24
is
which
clearly less than 9* during the adjustment period.
is
model gradualism helps because, through a subtle
In this specific
average and marginal costs,
in
can reduce the unemployment-cost of creation along the best
it
But gradualism by
feasible path.
other words,
it
between
difference
itself
cannot eliminate unemployment altogether.
Put
not an effective synchronizer of creation and destruction, because
is
requires reducing the former in order to reduce the latter. If
it
we could reduce unemployment
by reducing destruction independently, the optimal policy on the creation margin would be
to accelerate creation
up to the
efficient rate.
In more general models, with creation costs cq
M
tivity distribution in the
we have analyzed
when gradualism
is
and a non-generic
one can show that adjustment
sector,
the two extremes
+ c\H
in this section.
There
will
be a mixture of
is.
This highlights the
limitation of using a single policy tool to address distortions in two margins,
the need for the dual policy tools approach
we study
produc-
be phases of the transition
and others when "accelerationism"
beneficial,
will
initial
and points to
next.
Managed Adjustment
4
Even
if it is
does not
section,
the case that gradualism
mean
that other forms of
we turn
is,
at best, of limited use in restoring efficiency, that
managed adjustment cannot be more
effective.
In this
to the problem of constructing a set of policies that can restore efficiency
to the adjustment process.
Introducing Fractional Unemployment
4.1
Up
to
tion,
now we have
considered no role for unemployment in the efficient economy's transi-
mostly in order to sharpen our focus on the disruptions that result from appropriability
problems. This simplification
of zero
is
inconvenient
unemployment can be achieved
nite subsidies
frictional
and
taxes.
We
in
an
must therefore
unemployment by assuming that
when designing optimal
it
inefficient
start
policies, for
the limit
economy only by means of
infi-
by introducing some efficiency-role
facilitates reallocation
We modify our basic model by introducing a search cost C2(H/U)
for
during the transition.
that firms have to pay
per production unit created in addition to the previous investment cost (co
+ ciH).
Such
a cost can be derived from a standard constant returns Cobb-Douglas matching function
with equal
elasticities for
unemployment and vacancies, together with a constant vacancy
25
posting cost.
Since our concern
is
with inefficiencies that result from appropriability rather than exter-
we assume the standard
nalities in the search process,
bargaining share parameter
tion,
which
j3 is
equal to the unemployment elasticity of the matching func-
1/2 in our case (see,
is
efficiency condition that the worker's
e.g.,
Hosios 1990). This condition guarantees that the
thick-market and congestion externalities associated with search cancel out, and that the
decentralized search process
What
appropriation parameter should
ation costs?
appropriable.
0.5/(1
is efficient.
we apply
to the search-cost
Note that since search costs are incurred
21
From
— 0.5) =
1.
(11), this entails
The shadow
w(t)
(29)
component of
before bargaining, they are fully
— (3) —
an appropriation parameter equal to 0/(1
prices in equilibrium conditions (5)-(ll) are
=
H{t)
b(co
™
now
given by
H{t)
+ Cl* H(t)) + C2
uwr*"
cre-
'urn
and
= (1 + 6)(po + Cl H(t)) + 2c2
c(t)
(30)
It is clear
from
rise to positive
(29) that, even in the efficient case b
unemployment
= 0,
as long as creation H(t)
|||.
the presence of search costs gives
is
positive.
Distorted Margins and Policy
4.2
The
distortions that result
from appropriability problems
in this
economy are
reflected in
the two shadow price signals, (30) and (29). Their social counterparts, the social
cost of creation
and
social
shadow wage, correspond to the
limit case
when
shadow
b goes to 0:
2
^)^(fi)
(31)
eS(()=c + c 1 ff(t) + 2o2
(32)
The
social
shadow wage
positive because
20
.
is
^.
given by the social value of an unemployed worker, which
unemployment reduces search
costs.
What
is
is
now
the direction of distortion in
and Hammour (1994).
the nature of their timing, search costs are a prime example of appropriability, although their
dependence on unemployment and their relatively small empirical magnitude make them a rather special
case that we did not choose to emphasize.
See, e.g., Caballero
By
26
those two signals?
It is
w(t)
It is
easy to see that both are inflated in the presence of appropriability:
> w s (t)
and
no surprising, then, that the
>
c(t)
c
s (t),
economy exhibits
inefficient
b
for
>
0.
sluggish creation
and excessive
destruction following reform.
What
policies
can
and improve allocational
the private
shadow wage
economy? Along the destruction margin, since
efficiency in this
is
excessively high
job protection measures in sector
or
on each of the destruction and creation margins
offset the distortions
M may
and the economy
—
the efficient exit condition
to the
economy
in sector
X
for
it
is
shedding labor,
effective
shadow wage
in exit condition (6)
from (29) and (31) that by selecting
e(t). It is clear
e(t)
inefficiently
help. In particular, let e(t) denote a production
employment subsidy to sector M, so that the
becomes w(t)
is
= w(t) - w s {t) =
6(co
+ ciH(t))
restored. However, such
H(t)
u(ty
a measure cannot restore
full efficiency
depresses creation by raising the opportunity cost of labor employed
(see free-entry condition 5).
Turning to the creation margin, the excessively high
by providing sector
X
with a creation incentive
i(t),
effective creation costs
i(t)
The drawback
is
offset
so effective creation costs after the
subsidy in the free-entry condition (5) become c(t)—i(t). 22 In isolation
this restores the efficient entry condition
can be
(i.e.
when
e(t)
=
0),
when
= c(t) - cs (t) = b{co + CiiT(t)).
that by increasing hiring, creation incentives build up additional pressure
on the shadow wage
and therefore
an already high
level of
job destruction
should be clear that a combination of job-protection measures ec in sector
M and creation
(29),
intensify
in the contracting sector.
4.3
It
Optimal Policies
incentives
ic
in sector
X
is likely
to
remedy the economy's
We could assume the creation incentive
assumes the former.
i(t)
ills.
It is
straightforward to show,
to be either appropriable or not.
27
Our formulation
implicitly
by replacing those incentives into the entry and
exit conditions, that efficient
and
inefficient
equilibrium conditions coincide for the following canonical policies
e c (t)
=
w(t)
- w s {t)
and
e c (s)exp(-r(s-t))ds.
The
term
first
in the
while the second one
be protected
—
optimal creation incentive
solves the direct appropriability problem,
— only positive while sector M employment
offsets the increase in labor costs
As an
illustration, figure 2 simulates
are at their
positive
and needs to
due to the job protection measure
maximum
an example of those canonical
at the outset of the transition,
They then
costs of excessive destruction are highest.
down and unemployed workers
The
canonical policies e c (t) and
when
in
to incorporate such incentives in a
new productive
The
program that takes into account
plications for the resulting capital-labor ratio.
And
employment- protection
burden, or ongoing macroeconomic stabilization
geted incentives.
and
institutional
environment. Consideration should be
between encouraging capital investment or job creation,
e.g.,
structure.
art of the practitioner
given to the distortionary effects of implementing and financing the program.
—
Policies
capture the essential elements of the incentive
political constraints as well as the existing policy
incorporate elements
24
decline monotonically as reallocation
are incorporated into the
i c (t)
policies.
the need for creation and the
structure that a managed-adjustment program should target.
is
is
M. 23
sector
slows
ic
for
The
choice
example, has important im-
the existing policy environment
legislation,
a heavy corporate
— that either reinforce or
may
fiscal
offset the tar-
25
23
It may be puzzling that, in a two-sector model such as ours, the effect of our dual policy approach
extends beyond the relative subsidy it offers to the
sector. To answer this question, one must think of
investment and hiring out of U
which can be thought of as part of the informal sector
as originating
from a third, "reallocation" sector, at the expense of which the
and
sectors are encouraged.
X
—
24
L~
The
figure uses the following parameters: r
=
0.06, 8'
= L^ = 0.5.
=
1.2, 6
—
M
X
=
1,
Co
=
0, c\
=
1.5, C2
=
1.5,
and
Simulation details are available from the authors upon request.
Wage controls, for example, may be in place as part of a heterodox stabilization program, but they may
also help in reducing the bite of appropriability over the transition, when full-steam creation is necessary.
28
FIGURE
2:
OPTIMAL CANONICAL POLICIES
1.
0.9
0.8
0.7
/ment 0.6
5
emplcr
0.4
—u
CD
d
d
d
a
J
-1
*
^
^^
^
--....
2
3
time
Concluding Remarks
5
of appropriability arises naturally whenever different parties join into a
The problem
undertaking where the contribution of some
mon
is
specific.
When
those contributions are
not perfectly synchronized, the possibility of ex-post opportunistic behavior
tracts are precisely
their completeness
The
mechanisms by which the problem of synchronization
is
though joint-ownership may be a solution,
capital
when
labor
is
is
concerned.
The
it is
ground
Con-
arises.
managed, but
more a methodological benchmark than a description of
joint use of factors in production provides fertile
com-
reality.
for appropriability.
prevented by the inalienability of
resulting transactional difficulties
and labor have disruptive consequences on the economy. The problem
is
Al-
human
between capital
particularly acute
during times of adjustment, when a broad range of new investments are needed and new
contracts must be entered into.
the microeconomic
level,
When
appropriability problems cannot be overcome at
they receive a highly inefficient macroeconomic "solution" in the
form of depressed investment and thus more
By
ers.
this
difficult
employment opportunities
for
work-
restraining the bargaining position of workers in their transactions with firms,
endogenous response allows the economic system to guarantee the required return on
investment.
Appropriability results in
many of the features thought to characterize adjustment crises.
Massive job destruction in the contracting sectors
in the
expanding ones.
results in
is
accompanied by only a timid response
This lack of balance between creation and destruction not only
wasted resources, but also prolongs the transition unnecessarily. Since appropri-
ability leads to distortions in opposite directions
on the creation and destruction margins,
a gradualist policy approach can only relieve excess destruction by further depressing creation.
If
either.
An
expanding
it
is
not gradualist, our policy recommendation does not stop at laissez-faire
adequate policy response ought to combine vigorous creation incentives in the
sectors,
with job-protection measures in the contracting ones. This combination
of incentives was, to
some
extent, present in the activist restructuring path followed by
South Korea in the 1961-73 period
— with
its
aggressive promotion of investment in the
export sector combined with classic protection of domestic markets
— and has arguably contributed to
26
It
may be
its
success.
(see, e.g.,
Amsden
1989)
26
argued -that the key to Korea's success is flexible labor markets rather than intervention
The fact of the matter is that there was intervention, consisting of protection of the con-
(Krueger, 1983).
and various creation incentives in the export sectors (manufacturing).
Intervention also affected the labor market, where the government took an active role in the wage-setting
tracting sectors (primarily agriculture)
29
A
Appendix
In this appendix
Wn
(0) if
we compute
no adjustment takes
place;
(it)
adjustment; and (Hi) the welfare loss
(i)
initial
is
The
value of
Wn
9
(Q)
—
W
n
(0)
W
c
(0)
(i)
the value of welfare
-
Wn
(0)
from cold-turkey
from gradualist adjustment.
L = L~ + L~ between
full-employment allocation of labor
the
M and the X sectors
preserved and no investment takes place:
*-.<«,
(ii) If
(t)
=
6*Lx (t)
w(t)U(t)/bco,
—
To compute
*k _
^l Lm-
Lm (t) =
.
for all t
L x (t) — L —
>
0,
U(t),
and the flow of
and
(10) implies
write the present value of this flow as
e
^)
welfare under cold-turkey adjustment,
we
U(t)e-
rt
dt.
substitute expressions (12)
and
(15)
and U(t):
W
Subtracting
Wn
(0)
from
W
W
(33)
.
= -l±-j~(e* +
W(0)
(Hi)
°^
coH(t). Since (7) implies
we can
(33)
for w(t)
.
adjustment does take place, we have
welfare becomes
H
W
the welfare gain
2
obtained from the definition of welfare (20) assuming that the
is
(0)
model of section
for the linear
c
c
(0)
= ^-co(l + 6)L-.
r
(0) yields the
c
(0)
- W"(0) =
gain from cold-turkey adjustment:
(B*
- 1 - rco(l + 6))
Turning to welfare under gradualism, note that, by
—
.
r
(18), the integrand in expression
can be written as
fa +
—)
U(t)e-
H=
(0*
- 6(t))U(t)e- rt + ^y^(0(t) - rccM*)^.
move ahead of productivity and, at the same time, that workers
shared in the fruits of higher productivity (e.g., Mazumdar, 1994). Moreover, the large negative response of
employment during the-stabilization episode of the early eighties suggests that Korean labor markets may
not be that flexible. (This effect on employment was reflected only mildly in official unemployment figures,
mostly because of the dramatic fall in the participation rate. See Castaneda and Park, 1992).
process to ensure that wages did not
30
Replacing this into
W
where
9
Uc
spectively.
(33),
we
get
(0)-W c (0) = -fo°°(0*-0 9 (t))U 9 (t) e - rt dt
-^
and
U9
Jo°° [(**(*)
~
rco) U°(t)
-
(0*
-
rco) U%t)}
e~ rt
dt,
denote unemployment under cold-turkey and gradualist adjustment,
We can show that
the second term in the above expression
is
re-
negative by noting
that
(9(t)-rco)Z$-
=w(t)^+rcoU(t)
by
(18);
= co (H(t) + rU(t))
= -coe^Jmf^l
by
(10);
by (14).
rt
Multiplying both sides by e~ and integrating over time, yields
jH (0(t) - rco) ^p-e-
rt
dt
Since the value of above expression
paths of
9(t),
is
the second term in (34)
W
9
(0)
-
W
c
(0)
=
-co
f ^f^
dt
=
coU(0)
=
coL~.
constant for both the cold-turkey and the gradualist
is
=-
equal to zero. Thus,
f°° (0*
Jo
31
- &>(t)) U9 {t)e~ Tt dt.
References
[1]
Agenor, Pierre- Richard, and Joshua Aizenman (1994): "Macroeconomic Adjustment
with Segmented Labor Markets,"
[2]
Aghion, Philippe and Olivier
tral
[4]
9,
Blanchard (1994): "On the Speed of Transition in Cen-
NBER
eds.,
Macroeconomics
283-320.
Amsden, Alice H.
New
Working Paper No. 4769, June.
Europe," in Stanley Fischer and Julio Rotemberg,
Annual,
[3]
J.
NBER
(1989): Asia's Next Giant: South
Korea and Late Industrialization,
York: Oxford University Press.
Atkeson, Andrew, and Patrick
J.
Kehoe
(1992):
"Social Insurance
and Transition,"
manuscript, Chicago.
[5]
Caballero, Ricardo
J.,
and
Mohamad
ciency of Creative Destruction,"
[6]
Castafieda, Tarsicio, and
Labor Market,"
L.
NBER
Funkoo Park
Hammour
the Timing and Effi-
Working Paper No. 4768, June.
(1992): "Structural
Corbo and Sang-Mok Suh,
in Vittorio
"On
(1994):
Adjustment and the Role of
eds., Structural
Adjustment in
a Newly Industrialized Country: The Korean Experience, Baltimore: Johns Hopkins
University Press.
[7]
Chadha, Bankim, Fabrizio
ing,
Coricelli,
Unemployment, and Growth
and Kornelia Krajnyak: "Economic Restructur-
in a Transition
Economy,"
IMF
Staff Papers, 40-2,
744-780.
[8]
Coase, Ronald (1937): "The Nature of the Firm," Economica,
[9]
Corbo, Vittorio, and
J.
4,
386-405.
Miguel Sanchez (1992): "Ajuste de Empresas del Sector In-
dustrial en Chile durante 1974-1982," Coleccion Estudios Cieplan, vol. 35.
[10]
Cox-Edwards, Alejandra (1993): "Labor Market Legislation
Caribbean," World
in
Latin America and the
Bank Latin America and the Caribbean Technical Department
Regional Studies Program Report No. 31.
On
[11]
Diamond, Peter
[12]
Edwards, Sebastian (1988): "Terms of Trade,
in
(1994):
Time, Cambridge: Cambridge University Press.
Tariffs,
and Labor Market Adjustment
Developing Countries," World Bank Economic Review,
32
2,
165-185.
MIT LIBRARIES
milium""
IlllllV
3 9080
and Sweder van Winjbergen
[13]
009401149
"Disequilibrium and Structural Adjust-
(1989):
ment," in H. Chenery and T. N. Srinivasan, eds., Handbook of Development Economics,
vol. 2,
[14]
[15]
Amsterdam: North-Holland.
Fallon, Peter R.,
and Luis A. Riveros
The World Bank
Policy, Planning,
(1989):
"Adjustment and the Labor Market,"
and Research Working Papers, No. 214, June.
Gavin, Michael (1993): "Unemployment and the Economics of Gradualist Policy Reform," Columbia University, typescript.
[16]
Grout, Paul A. (1984): "Investment and Wages in the Absence of Binding Contracts:
A
[17]
Nash Bargaining Approach," Econometrica,
Hart, Oliver, and John
Moore
52-2, 449-460.
"Incomplete Contracts and Renegotiation,"
(1988):
Econometrica, 56, 755-786.
[18]
Horton, Susan, Ravi Kanbur, and Dipak Mazumdar, eds. (1994): Labor Markets in an
Era of Adjustment, Washington: The World Bank.
[19]
Hosios, Arthur J. (1990):
"On the
Efficiency of Matching
and Unemployment," Review of Economic
[20]
Klein, Benjamin, Robert G. Crawford,
gration, Appropriable Rents,
Law and Economics,
[21]
Krueger,
Anne O.
and Related Models of Search
Studies, 57, 279-980.
and Armen A. Alchian (1978): "Vertical
Inte-
and the Competitive Contracting Process," Journal of
21, 297-326.
(1983): Trade
and Employment in Developing Countries, Chicago:
University of Chicago Press.
[22]
Lapan, H. E. (1976): "International Trade, Factor Market Distortions, and the Optimal
Dynamic
[23]
Policy,"
American Economic Review, June.
"Entry-Exit, Learning, and Productivity Change:
Liu, Lili (1993):
Chile," Journal of
[24]
Lopez,
Ramon
Labor Market
E.,
in
Development Economics,
and Luis Riveros
Evidence from
42, 217-242.
(1989):
"Macroeconomic Adjustment and the
Four Latin American Countries," The World Bank Policy, Planning,
and Research Working Papers, No. 335.
[25]
Marshall, Isabel (1992):
"Liberalizcion Comercial en Chile
y su Impacto sobre
Eficiencia Tecnica Industrial: 1971-1986," Coleccion Estudios Cieplan, vol. 35.
33
la
[26]
Mazumdar, Dipak
and Dipak Mazumdar,
ton:
[27]
"The Republic of Korea,"
(1994):
eds.,
in
Susan Horton, Ravi Kanbur,
Labor Markets in an Era of Adjustment, Vol.
2,
Washing-
The World Bank.
Mussa, Michael (1978):
"Dynamic Adjustment
in the
Heckscher-Ohlin-Samuelson
Model," Journal of Political Economy, 86-5.
[28]
Neary,
J. P.
(1982):
ment Assistance,"
Wage
"Capital Mobility,
in J.
Bhagwati,
ed.,
Stickiness,
and the Case
for Adjust-
Import Competition and Response, Chicago:
University of Chicago Press.
[29]
Roberts,
Mark
Performance in
[30]
and James R. Tybout,
Producer Heterogeneity and
the Semi- Industrialized Countries, unpublished.
mimeo
April.
Thomas, Jonathan, and Tim Worrall
of Expropriation," Review of
[32]
eds. (1994):
Shimer, Robert (1995): "Microfoundations of the Optimal Speed of Transition,"
MIT,
[31]
J.,
(1994): "Foreign Direct Investment
Economic
and the Risk
Studies, 61, 81-108.
Tybout, James R., Jaime de Melo, and Vittorio Corbo (1991): "The Effects of Trade
Reforms on Scale and Technical
Efficiency:
New
Evidence from Chile," Journal of
International Economics, 31, 231-250.
[33]
Williamson, Oliver E. (1979): "Transaction-Cost Economics: The Governance of Contractual Relations," Journal of
[34]
(1985):
The Economic
Law and Economics,
Institutions of Capitalism,
34
2554
;27
22, 233-261.
New
York: Free Press.
Date Due
AU& $
i
rase
Lib-26-67