Digitized by the Internet Archive
in
2011 with funding from
Boston Library Consortium IVIember Libraries
http://www.archive.org/details/onspeedoftransitOOaghi
^
rA
working paper
department
of economics
On the Speed of Transition in
Central Europe
Philippe Aghion
Olivier Jean Blanchard
No
93 -8
Jiine
1993
massachusetts
institute of
technology
50 memorial drive
Cambridge, mass. 02139
1
On the Speed of Transition in
Central Europe
Philippe Aghion
Olivier Jean Blanchard
No.
93-8
June 1993
AUG
1 2 1993^
RECEIVE
I
On
the Speed of transition in Central Europe
Philippe Aghion
June
1,
Olivier Jean Blanchard
*
1993
Abstract
Transirion in Central Europe involves the closing and restructuring of state firms, as
well as the
and the
emergence of a new private sector. The speed of closing and restructuring
rate of private job crearion
unemployment
new
It
in
private jobs.
cum
affects
Our paper
determine the dynamics of unemployment.
And
both the decisions to restructure as well as to create
presents a model which captures these interactions.
characterizes the positive and normative properties of the equilibrium speed of
transition
and unemployment
rate,
and the
role of policy.
Analytics of Transition
In
system
is
all
Central European countries, the legacy of the previous economic
a large
and
must
parts of firms,
ailing state sector.
made
ding before they can be
been extremely
The
close.
Labor productivity
is
low.
Many
firms, or
others need drasric restructuring and labor shed-
to prosper. But, so
far,
the pace of restructuring has
slow.
The evolution
of the state sector accounts however only for one side of the tran-
sition process.
The
been impressive so
ing
many
other
far,
is
but
new
the growth of a
it
private sector.
The growth
has not taken place across the board.
obstacles, ft-om poor financial markets to insufficient
And
it is
has
fac-
human capital and
expertise.
This transition process
speed of transirion
fast
?
raises
many important
How and why can
it
derail
and what are those of going too slow
These
are the issues
we take up
ing two sectors, the state
one of the two sectors, or unemployed.
one
in
which
all
"What are the
determining the
risks of
going too
try to affect
?
in this paper.
and the private
?
is
Should the government
?
the speed, and through which instruments
What
issues.
We
sector.
think of the
economy
as hav-
Workers are either employed
in
We think of the pre-transirion economy as
workers work in the state
sector.
We
then formalize transirion as
the result of two interacting processes, the restructuring of state finns
hand, and the growth of new private firms on the other.
The
on the one
rate of restructuring
determines the flow into unemployment; the rate of private job creation deter-
mines the flow out of unemployment. Together they determine the evolution of
unemployment and the speed of transition
We think that our contribution lies
ceptual structure as in the results
-but not
all- of
a role for the
as
we
^.
much
in offering a simple
and
flexible
obtain. But those can be stated simply.
con-
Most
our results point to too slow a speed of transition, and thus
government
ever to a clear risk that too
to accelerate the process.
fast
They
also point out
for
how-
a speed can lead to derailment and an ultimately
X
Analytics of Transition
unsuccessful transition.
Our paper
Section
1
is
organized as follows.
offers a
and are closed
at
benchmark model. In
some constant
flow out of unemployment
tion of both
of unemployment, while
state firms are low productivity firms
rate, leading to a flow into
comes from
wage and non wage
it,
private sector job creation,
costs.
non wage
unemployment. The
Wages
which
is
a frinc-
in turn are a decreasing frjnction
costs are an increasing fijnction of unemploy-
ment. By assuming that state firms simply close
at a
constant
rate,
the model does
justice neither to the possibility of restructuring, nor to the decisions of state firms
as to restructure or not. But
it
sheds light on the joint dynamics of unemployment
and private sector job creation. Those dynamics are characterized
and
3.
The main
unemployment,
lesson
kill
is
in Sections 2
that too high a rate of closing can, through
private sector growth
and
its
effect
on
derail the transition.
Sections 4 and 5 refine the treatment of the state sector. Section 4 allows for
restructuring rather than closing. Restructuring
firms
become
ers are laid
in effect,
private, productivity
off.
is
increased,
is
formalized as a process where
and a proportion of the work-
Section 5 focuses on the decision to restructure.
It
argues that
workers can block restructuring, so that restructuring only happens
workers approve
it.
It
derives the implications of such an assumption. Clearly, as
restructuring implies that
some workers
set as in private firms, restructuring
compared
if
to those in private firms,
is
will lose their job
less appealing,
and that wages
will
be
the higher state firm wages
and the higher unemployment. Thus, the
rate
of restructuring and the flow into unemployment are decreasing in the level of
unemployment.
Sections 6 to 8 put the pieces together. Section 6 characterizes the equilibrium
rate of
unemployment and speed of transition. Section
properties,
and the case
calibrations.
for
7 looks at their normative
government intervention. Section 8 presents rough
Analytics of TTansition
Section 9 concludes, and indicates what we see as some of the limitations of our
mention one from the
analysis. Let us
start.
We
carry our analysis under the as-
sumption of a hard budget constraint, that firms cannot act
obtain larger subsidies from the government.
true so far in the major Central
of our paper). But the constraint
may
well not hold in the future.
Our
The economy
hence the
be tested again and again, and
only excuse for not modelling
it is
that this
far.
composed of two
is
sectors, the state sector with
the private sector with employment
E+N+ U =
that
We
rate.
so that
wages,
E =
first
1
where
and
N
in the state
U
labor force
is
normalized to one, so
unemployment, equivalently the unemployment
is
= U =
N. The
employment E,
transition,
0.
and then
We
all
workers work in the state sector,
describe the behavior of
employment and
in the private sector.
state sector
Both price
many
1,
assume that before the
The
1.1
liberalization
and the demise of the CMEA had the effect of making
jobs unproductive from the start of the transition process.
by assumirig that, as transition
We
likely to
in Russia;
A benchmark model
1
(1)
is
to
see this as having been mostly
European countries (but not
title
would have taken us too
We
such a way as
in
starts,
employment
We
capture this
in the state sector drops to
assume that the remaining workers, Eq, have constant marginal product,
The low
productivity of state firms
is
captured by assuming that x
is
less
than
x.
y,
the marginal product of workers in the private sector, described below.
(2)
Three years into the
state
owned. But,
in the
transition,
most state firms have remained nominally
absence of state supervision and control, firms have acted
Analytics of Transition
largely in the interest of their workers, with a fairly short horizon.
firms,
Thus,
in
workers have increased wages so as to appropriate quasi rents.
We capture
this
by assuming that the wage in state firms
We —
where a
is
clx
—
is
equal
to:
z
(1.1)
greater than one, reflecting the appropriation of quasi rents by workers,
and z are taxes per worker. Once the workers have appropriated
incidence of taxes levied on the firm must
(3) Finally,
we assume
fall
entirely
all
rents, the
on wages.
that the speed of closing in the state sector
is
given by
dEjdt = -s
We
shall start
by thinking of s as an exogenous parameter, say under the control
showing the basic dynamics of the model most
starkly. But, later,
assumption in two ways. First we shall allow
parts of firms,
s:
(1.2)
of the government. Starting with the assumption of an exogenous 5
this
many
can be saved and restructured,
for the fact that
we
is
useful in
shall relax
some
firms, or
albeit at the cost of reductions in
employment. Second, and more importantly, we
shall
endogenize
s,
by looking at
the choice of state firms/workers between restructuring or not restructuring.
The
1.2
(1)
The
private sector
private sector has
overnight.
We
grown
fast,
assume private job creation to be given
dN/dt =
where y
wage
is
but clearly cannot replace the state sector
by:
a{y-z-w)
the constant marginal product of labor in the private sector,
in the private sector,
(1.3)
w
is
the
and z are again taxes per worker, which are assumed
Analytics of Transition
be the same
to
and
for private
Private job creation
ployment and the
is
thus characterized by a relation between the change in em-
is
new
capacity not exceed retained earnings
"'.
The
indeed that that new private firms do not use outside finance. Or we
can think of
tional,
This relation can be seen as coming ft-om the
profit per worker.
constraint that investment in
evidence
state firms.
as
it
coming
and some coming
fi^om costs of adjustment,
such aspects
ft^om
some
physical
as learning by doing,
and conven-
accumulation of
information, development of reputation in goods and financial markets and so
"•.
If
comes from
the relation
costs of adjustment,
it is
likely
however
looking: before creating jobs, firms will worry about both current
pected
We
profits.
ignore this aspect for the time being, but take
on
to
be forward
and
fijture ex-
it
up in Section
3 below.
(2)
The evidence on
private sector
idea that the private sector
reasonable.
We
6
hires to
is
unemployment
unemployment
is
later to derive
it
explicitly:
^.
Assume
H/U.
the interest rate, and
H/U
from efficiency wage considerations.
fijrther that there
are equal to
Under those assumptions,
equal to
is
(1.4)
is
the ratio of
a constant.
It is
useful for
be the values of being unemployed and employed in the private
sector respectively.
H
+ c{r+H/U)
the exit rate ft-om unemployment, and c
easily derived
Vu and VV
b
benefits, r
is
so that hires,
sketchy, but the
take private sector wages to be determined by:
This equation
Let
is still
wage depends on labor market conditions appears
w=
where
wage determination
is
dN/dt, and that
no turnover
all
in the private sector,
hires are firom
the probability of being hired
unemployment
when unemployed
is
Analytics of Transition
The two
values thus follow "arbitrage" equations:
r
'
"^
''
i
When
no
risk
forever
Many
=
b
rV^
=
w + dV^/dt
+ {H/U){Vrsf-Vu) + dVu/dt
(1.5)
(1.6)
unemployed, a worker receives unemployment benefits
H/U
bility
rVu
of being employed.
When employed,
and has proba-
6,
a worker faces, by assumption,
of becoming unemployed again, and thus receives the private sector wage
after.
wage considerations can be summarized by the condition that
efficiency
wage such
firms choose a
that the value of being
employed exceeds the value of
being unemployed by some amount, thus such that Vjv
that assumption,
which obviously implies that dV^ jdt
—
Vfj
=
c,
Note
that, in that equation, higher
the wage one for one.
unemployment
rate
Note
itself,
also that the
wage
is
unemployment
1.3
to workers,
and
affects the
as well,
an increasing function not of the
se,
It is
clearly
which
is
of
wage.
Taxes and unemployment benefits
The remaining assumption
efits.
which
Under
wage equation
but of the exit rate out of unemployment.
thus,
0.
benefits increase
the probability of getting a job, not the unemployment rate per
concern
>
— dVu/dt =
taking the difference between the two equations above gives the
(1.4) above.
c
We
relates taxes,
unemployment and unemployment ben-
assume that taxes are levied equally on employment,
to finance
unemployment
benefits.
state
and
private,
This implies:
Ub = {l-U)z
(1.7)
This relation has a straightforward implication. Higher unemployment, given
unemployment
benefits, leads to higher taxes per worker; thus, ceteris paribus,
Analytics of Transition
higher unemployment decreases private job creation.
There
obviously more to the relation between the process of transition and
is
balance than this simple relation.
different channels through
We see
this formalization as a
which a depressed economy
and thus the growth of the new private
itability
metaphor
These may be
sector.
been
flow of retirements, clearly as an alternative to
unemployment
demand
for the
output of private
unemployment decreases support
form process
and
uncertainty,
There
firms
is
will
go on;
for
for older workers;
may be through
may mean lower aggregate demand, and
firms.
One can
think of yet others: higher
reform and thus the likelihood that the
this in turn leads to
lower expected
profitability,
just
It is
that the tax
burden on private
One may easily
channels through which the burden may depend on that composirion,
if
higher
on unemployment, and does not depend on the composition
of employment between the private and the state sector.
ple
re-
in turn lower private job creation.
a strong assumption implicit in (1.7).
depends
other, but
Poland an unusually large
in
those higher retirement benefits also need to be financed. These
lower
for the
affects adversely the prof-
related, fiscal channels: for example, there has
aggregate demand: higher unemployment
fiscal
state firms receive subsidies
and private firms do
not, or
if
think of
for
exam-
state firms
and
private firms are taxed differently. Allowing for such distriburion effects would,
complicate the analytics below; we do not think that
it
would change very much
the substance of our results ^
1.4
laking care of comers
In specifying our assumptions,
it
might have been more natural to specify the
equations for state and private employment in terms of rates of change rather than
changes,
dN/dt.
in terms of
i.e.
We
{dE/dt)/E and {dN/dt)/N than
have chosen our specification because
analysis later, leaving only
employment
in
one
state variable,
it
in terms of
dE/dt and
substantially simplifies the
unemployment, rather than two,
one sector and unemployment. But,
as a result of these
assump-
Analytics of Transition
tions, the
economy can
are reached.
and we have
Those conditions only
similar condition
A''
=
0.
to
must hold
make
if it is
£ >
for private
what happens when those
end of the
play a role at the
employment can only decline
the condition (1.2) only holds for
innocuous
to state
economy go very wrong, but they must be
things in the
First, state
hit corners,
first
= 0,dE/dt =
employment. Here,
it is
a slightly different assumption, that for {y
Namely, private firms can close
if
when
specified nevertheless:
positive in the
0: for i^
transition or
they are losing money.
place, so that
0.
Second, a
convenient and
—
z
— w) <
We shall
0,
ignore
these conditions in the text, (but not in the derivatior^s and graphs which underlie
the text) only mentioning
,
them when
Unemployment and
2
In our
directly relevant.
the speed of transition
benchmark model, the speed of closing determines the inflow
into
unem-
ployment. Unemployment in turn determines private job creation, the out-flow
firom
unemployment. Together, these two relations thus determine unemployment
dynamics.
Working backwards, we
first
derive the relation between
unemployment and
pri-
vate job creation. Solving for the wage, using equations (1.4), (1.3) and (1.7)
gives:
iw-b) =
Replacing
[ca/iU
w by its value
from
+
ca)][y
(2. 1),
+
{r/a)U - ((1/(1 - U))b]
and z by
its
(2.1)
value from (1.7) gives private job
creation as a function of unemployment:
dN/dt = a[U/{U +
Unemployment
channel
is
ca)][y
-re-
affects private job creation
captured in the
first
(1/(1
-
U))b]
= f{U)
(2.2)
through two channels. The direct
term: the higher
is
unemployment, the lower
is
n ^^fv^
/^,
I
^
u
^\^i.Mf
^
£-,Ay
u
Analytics of TTansition
10
the wage, and thus the higher
The second term
private sector job creation.
is
captures the effect of unemployment through contributions, taking into account
on the
implicitly the incidence
The
higher
the lower
How
is
is
private sector wage,
which partly
offsets the effect.
unemployment, the higher are unemployment contributions, thus
private sector job creation.
these two relations determine
unemployment
is
shown
in figures
and
I
2,
which, each, plot private job creation and state job closing as functions of U.
In both, private job creation
origin.
When
private
employment
exit rate
unemployment
and thus
dominates
equal to zero, the wage
is
would lead
to
an
infinite
wages. As unemployment increases, the effect which
increases.
unemployment on wages, so
As unemployment
private
employment growth
that private
gets sufficiently large however,
declines.
effect of
unemploy-
As unemployment
gets
enough, wages and contriburions exceed the marginal product of labor, lead-
ing to the disappearance of the private sector. Thus, at low
unemployment
more job
leads to
ployment leads to
creation; at high
unemployment, higher
unemployment, higher unem-
job crearion.
less
In both, the decline in state
is,
locus starts from the
so high as to prevent
unemployment on contributions dominates the
ment on wages, and
large
is
dN/dt = f(U). The
the direct effect of
is
employment growth
given by
crearion; any positive rate of hiring
infinite
initially
the effect of
is
employment on the other hand
is
given by
5,
which
by assumption, constant and thus independent of unemployment, a horizontal
line at s.
Figures
1
In figure
3,
so that
starts
and 2 consider the two configurations which these equations can
1
,
the
yield:
maximum rate of private sector job creation is such that it exceeds
unemployment
anywhere between
during the transition.
ment decreases
If
to Ui.
at that point
is
decreasing. In that case,
and Ui unemployment goes
,
the
economy
starts
to Ui
if
the
economy
and then stays there
between Ui and U2, then unemploy-
Beyond U2, the weight of contriburions
is
too large; the
Analytics of Transition
1 1
enough and unemployment grows
private sector does not take off fast
until the
private sector actually closes. To the right of point A, the private sector never
takes
off.
In figure
the
2,
employment
maximum
is still
rate of private sector
initially,
employment
is
employment
decline, the fiscal effects
employed
There
is
down. Unemployment
one simple lesson from
If
the speed
change of private sector
employment growth can
become dominant, and
private
relentlessly increases until
this
It is
model: there
enough
is
a
offset state
employment
everybody
maximum
that value of 5 such that the
higher, the rise in
is
is
un-
two
speed at which
loci are exactly
unemployment never allows the
private
to replace the jobs lost in the state sector.
earlier, private
part, forward looking.
the
rate of
A forward looking private sector
As we pointed out
will
instead such that un-
^.
sector to grow fast
3
and the
increasing. But, before private
the state sector can be closed.
tangent.
is
increasing at that point. In that case, starring from low un-
employment, things look good
starts slowing
change
not invest
fijture.
if
Many
private firms,
likely to be, at least in
is
and especially foreign
direct investors,
they expect conditions to deteriorate, and profits to shrink in
But in turn,
profits in the fijture
may be low
if
private job creation
is
unemployment.
insufficient to avoid rising
To explore
sector job creation
this interacrion,
we modify our description of private employment
cre-
ation, equation (1.3), to read:
where
dN/dt
=
gtV
(3.1)
rV
=
{y-z-w) + dV/dt
(3.2)
V is the value of a new private job. Job creation now depends on the value
^,qcYt 3
Vr
23
U rO
2
Analytics of Transition
new
of a
that
costs, z
The value
job.
of a
new
job follows an arbitrage equation,
+
w. Such a relation
is
easily derived
from quadratic costs of adjustment.
Under these assumptions, the economy reduces
dU/dt
=
s-arV
rV
=
f{U)
where, as before, f{U)
The
locus {dU/dt
=
0)
= [U/iU +
is
locus (dV/dt
erties as the
is
zero,
f{U)
=
0)
all
and V:
-
(3.4)
U))b].
V — U space,
at
(F =
s/ar),
V =
f{U)/r and thus has the same propIt
V
goes through the origin
is
zero. It
it
downward sloping, and
is
as, if
upward sloping
increases
U
ini-
unemployment
V becomes equal to zero at
b/y.
There are again two
We
U
equal to state employment destruction.
the rents and thus
it is
in
ca)]dV/dt
unemployment decreases wages more than
contributions. Eventually,
U = 1-
(1/(1
locus characterized earlier.
wages extract
tially as
given by
-
line in the
is
two equations,
(3.3)
ca)][y
an horizontal
is
to
+ [U/{U +
such that private employment creation
The
which implies
the present value of the difference between y and direct and indirect
is
it
1
cases,
depending on whether the two
loci intersect or not.
concentrate on the case where they intersect; the conditions for this are the
same
as before.
The dynamics
paths.
are characterized in figure 3.
There are always two equilibrium
The first path is given by A'B' (under myopic expectations,
by OB' instead). The private sector
to maintain
is
optimistic,
unemployment constant. The
path goes to the wrong place. Firms are
ment creation
goes from
is
it
would be given
and quickly creates enough jobs
transition
is
a
smooth one. The second
less optimistic, create less jobs,
never enough, and their fears are proven founded.
employ-
The economy
A to B and asymptotes to C as above *.
Thus, even
if
the speed
is
below the
maximum speed derived
in the previous sec-
Ancdy tics of Transition
tion,
expectations matter very much.
If
5
is
and indeed does
the policy lessons from this exercise
control
There
no obvious sense -at
is
speed makes the good outcome more
nounce
a policy
?
Suppose that the government can
path
will
result.
same purpose. But one senses
benefits
is
more
rate higher
Under such an announcement, the only
One can
that there
is
a substantive difference
explicit treatment of the
armouncement of lower
benefits
serves the
between the
government's ob-
when unemployment
problems of credibility and time consistency.
likely to suffer firom
Thus, within our model, the lesson
is
that, in order to avoid self-fulfilling forecasts
of the failure of transition, for example, by foreign investors, the government
want
to
announce
unemployment,
it
namely armounce
a flexible policy,
will
slow
per-
think of other announcements
unemployment exceeds Ui formally
if
two annoucements. Even without an
clear that the
an unemployment
For example, the announcement that unemploy-
benefits will be eliminated
it is
which a lower
But the government can however an-
likely.
be the good path.
which achieve the same
jectives,
in
which eliminates the bad outcome. Any policy which announces
than, say Ui, achieves this result.
ment
we can see-
least that
that the speed of closing will be equal to zero for
fect foresight
anticipated to
it is
fail.
What are
s.
3
constant, there are always two self
one where transition succeeds, and one where
fulfilling paths,
fail
1
down
that, conditional
may
on high
the process of closirig state firms. In doing so,
the government will avoid the emergence of high unemployment, and therefore
will
not have to actually decrease
How
robust
risky to
is
this result
go slow as
it is
?
to
It
go
s.
goes against the often expressed view that
fast.
One
possible rationale for that view
with time, state firms may increasingly ask
private firms.
Our model
very
much
relies
erate under a hard budget constraint.
in Russia,
far.
it
for
and obtain
subsidies,
on the assumption
While
this
assumption
more
it is
is
as
that,
so than
that state firms opis
clearly incorrect
has been largely satisfied in the major Central European countries so
Analytics of Transition
4
14
Restructuring instead of closing
The approach
been
so far has
from the new private
sector.
assume that
to
all
The evidence however
wrong, that a number of state firms are
likely to
must eventually come
activity
suggests that this view
is
be profitable, although not with-
out closing of some activities, and not without considerable labor shedding.
This possibility of transformation can be introduced easily -perhaps too easilyin the following way. Reinterpret the initial
lows. Initially, there are
employing each
1
Eq
worker
state "firms",
(a
fol-
each producing x units of output, and
convenient normalization). Over time, those firms
are transformed. Transformation
(1)
assumptions about state firms as
is
such
that:
each firm now employs only A workers, A
<
1,
and each worker produces
y,y > X units of output, so that each firm now produces Xy units of output,
which can be greater or
is
less
than
clearly just a special case of this
(2)
the
new
x. Closing, the case
we have considered
more general formulation, with A
firm operates from then
on
like a private firm, so that
set as in the rest of the private sector, at level
we
shall
when
now
far,
0.
wages are now
w.
Thus, we think in effect of transformation as restructuring
private sector will
=
so
cum privatization. The
include both transformed state firms and
new private firms;
convenient, refer to the second source, as "new" private job cre-
ation.
We shall maintain for the moment the assumption of an exogenous speed s, where
5
is
now
the speed of restructuring rather than of closing.
We shall
endogenize
it
in the next section.
When restructuring takes place
The equation
for the
at rate 5, (1
—
A)^ workers become unemployed.
behavior of unemployment becomes:
dU/dt =
s{l-X)-f{U)
(4.1)
5
Analy tics of Transition
The dynamics
and
1
(
1
2,
of
1
unemployment and
transition are thus the
same
as in figures
with the flow into unemployment being given by the horizontal line
— A)5 rather at
s.
The conclusions we derived earlier above still
at
hold, in slightly
modified form:
There
is
a
maximum
speed of restructuring -which
is
simply equal to the maxi-
mum speed of closing we derived earlier, multiplied by 1/(1 — A)- consistent with
a successful transformation. For speeds
ing directly increases output, that
much. This eventually
kills
is
beyond
even
this
Xy >
if
maximum, even
private sector job creation. Eventually, both the
For speeds lower than the
maximum, and
eventually converges to U\
,
to
£/i
0,
and
as long as
the lower of the two
stays there during the transition.
restructuring
,
is
restructur-
unemployment increases too
x,
and the transformed private sector become unprofitable and
dU/dt =
if
contributing both directly
increase in the private sector.
The
indirect effect
ployment, which, given that f'{Ui)
>
close.
Uq <
U-i,
unemployment
the
economy
rates at
When unemployment
(,Xs)
is
and
is
which
equal
indirectly {f{U)) to the
through the increased unem-
leads to faster job creation in the
0,
new
new
private sector as well.
When
forward looking behavior
is
introduced in the new private sector, as we
did in the previous section, then constant
announced speeds of restructuring
associated with two paths, one successful
and one unsuccessful. And,
nouncing that restructuring
will
be halted
if
unemployment
is
are
again, an-
too high can help
sustain the virtuous circle equilibrium.
5
We
Restructuring as a choice
have so
far
viewed the speed of restructuring as a given parameter, under
the control of the government.
Many of them
to operate at
We now view
indeed have the choice,
it
instead as a choice by state firms.
at least for
some
time,
between conrinuing
low levels of productivity, or restructuring, getting
rid of
redundant
6
Analytics of Tramition
1
When
operations, eliminating labor hoarding, and increasing productivity.
they take the decision to restructure
decisions mostly represent, as
If firms'
?
have argued, the interests of their workers, the answer must be that firms
when
restructured only
the workers agree to
What
it
much
relevance, lb see that,
ment
in the
means
for
"workers to agree"
let
untransformed state
unemployed.
And
let
we can
use the
is
will
we
be
it.
ambiguous, and
an issue of
this raises
represent the value for a worker of employ-
sector. Let, as before,
V)v be the value of being
which, under our assumptions,
sector so that
Ve
is
will
Vu
employed
be the value of being
in the transformed firm,
the same as that of working in the
same symbol. Recall
new
private
that restructuring implies that
a proportion A of the workers will remain employed after restructuring, and a pro-
portion
If all
1
- A
will
become unemployed.
workers perceive an equal probability of keeping their job after restructuring,
then -under the assumprion of risk neutrality, an assumption that we have made
implicitly until
now
in defining the various value funcrions as linear in
benefits-, the condirion for restructuring
XV^ +
{1
wages and
is:
- X)Vu > Ve
(5.1)
For restructuring to take place, the expected value after restructuring must be
greater than or equal to the value absent restructuring.
But, in most firms, the assumption that workers choose under a veil of ignorance
is
clearly wrong, in
surely
have to
the firm. Even
close.
if
most
cases,
it is
What happens
easy to idenrify parts of the firm which will
then depends on the decision process within
decisions to restructure require, as they do de facto in Poland,
quasi unarumity of the workers, the condition above
who know
they will lose their job can be
pay for example, by those
in
many
cases,
workers
who
fijlly
may still hold if those workers
compensated, through severance
are likely to keep theirs. But the evidence
who were
is
that,
likely to lose their job as a result of restructuring
7
Analytics of Transition
1
have blocked attempts
at
such restructuring. Under the extreme assumption of
unanimity and no- transfers, the condition
for restructuring
must
be:
Vu > Ve
This condition, that restructuring
as a result are as well off as
is
if
(5.2).
We shall
work
is
Endogenous
only
if
those
who
lose their jobs
somewhere between the two conditions,
what follows with condition
when working with
things are modified
6
in
will take place
they remained employed in the untransformed firm,
probably too strong, and reality
and
(5.2)
(5.1),
(5.1)
and indicate how
(5.2) instead.
unemployment and the speed of
restructuring,
transition
We now
put both parts of the model together, and start by giving a general de-
scription of the transition, with details
Starting from an
unemployment
rium unemployment
From then
rate,
qualifications to follow:
rate Uq, the
economy converges
U', such that the condition
economy
on, the
and
stays at this
unemployment
(5.1)
to
an
equilib-
holds with equality.
rate, until
the end of the
transition.
At
[/
=
[/*,
job creation in the
of restructuring
new
private sector
is
=
f{U'). Thus,
sector job creation, from transformation of state firms
is
equal
+
fiU')
=
U=
U', total private
and from new private sector
= f{U')/il-^)-
transition path, the value of being
given by V}v
at
unem-
to:
sX
Along the
The speed
thus given implicitly by the condition that inflows into
is
ployment equal outflows, or (1 - A)s*
creation
equal to f{U').
w/r. As the wage
is
employed
in the private sector
a decreasing fiinction of
is
unemployment, V}v
Analytics of Transition
is
J
a decreasing function of unemployment.
And
from the conditions Vyv
= AVV + (1 — ^)Vu, the three values
unemployment. And from the same conditions,
and Ve
sector are better off than those
off
employed
are
=
Vjy
8
+c
decreasing functions of
all
those employed in the private
in the state sector,
who
are in turn better
than the unemployed:
V^ > Ve = V^-(1-X)c > Vu =
6.1
V^-c
(6.1)
Characterizing U'
We now
characterize
absent restructuring
is
U'
formally.
of being employed in a state firm
characterized by:
rVE =
Along the
The value
transition path,
ax-z + dVE/dt
when U
is
(6.2)
equal to U*, unemployment benefits and
thus taxes per worker, z are constant. Thus,
dVE/dt =
0,
and Ve
is
given by
{ax — z)/r.
The
values of being unemployed and of being employed in the private sector re-
spectively are given by equations (1.5)
are also constant,
and dV^/dt
b
(1.6). If
= dVu/dt =
rVN = w =
rVu =
and
b
+
c{r
0.
[/" is
constant,
Thus, V/v and
Vu
+ f{U')/U')
+ U{U')IU'){Vn -
Replacing in equation (5.1) and noring that Vjv
=
Vu)
Vj/
thenw,f{U)/U
are given by:
(6.3)
(6.4)
+ c gives:
ax-z = b + cf{U')IV + Xcr
(6.5)
Analytics of TTansition
19
U' must be such
Equilibrium unemployment,
hand
sector, the left
are
unemployment
side of (6.5)
equal to the
,
and the
benefits,
make
as to
sum
the
wage
of three terms.
exit rate out of
The
two
first
unemployment times
difference in value between being employed in the private sector
employed. The third
in the state
c,
the
and being un-
the annuity value of being employed in the private sector
is
over the value of being unemployed, times the probability of being employed in
the restructured firm.
If
we had used condition
restructuring had to be approved by those
does the economy get to U'
ployment from the
restructuring,
initial
If,
?
shock, Uq,
\s
and a discrete increase
Uq exceeds U', there
unemployment
they were going to lose their
is
at the start of the transition, the
less
in
than U', there
unemployment
no restructuring
is
a discrete
U=
until
until private sector
U'.
unem-
amount of
If
instead,
growth has reduced
to U'.
To characterize U'
fijrther,
rewrite (6.5) using the expression for
balanced budget condition that z
= bU/{l -
Figure 4 plots both sides of the equation.
right
f{U) and the
U). This gives:
- cr) + _£_(6/(l jjf^Ay
[U + ca)
K^ + caj
ax - Xcr = A{U) =
Qi - Act-. The
who knew
is if
term would be absent.
jobs, this third
How
(5.2) instead of (5.1), that
hand side,
/!((/)
is
The
left
hand side
equal toy — cr for
is
an horizontal
U=
0. If
(6.6)
[/))
line at
y — cr — 6
>
0,
the condition that the marginal product of the private sector exceeds the lowest
wage the private sector may
however,
increases
it
pay,
and tends
A{U)
\s
initially
downward
to oo as f/ tends to one. Figure 4
depending on parameters, three configurations are
In the
first,
state workers just
state wages,
which
sloping. Eventually
do not want
shows
that,
possible:
to restructure,
and
prefer to keep their
in part capture quasi rents, to the risks of restructuring
associated unemployment. This will be the case
when A{0) is below
and
the horizontal
()(^-cAr
..
T)'^U^-C
i.
Afu.)
^
_
CCL
ii
+
.cr)
ca
20
Analytics of Transition
locus, or equivalently
ax — cXr >
if
to zero, the process of transition will
and zero unemployment
new
of the
—
cr. In that case,
never
start.
No
if
Uq
equal
initially
is
restructuring will take place,
high private wages, preventing the growth
will lead to
private sector as well.
y
If
unemployment
is
sector job creation will also be initially positive, but as
initially positive, private
unemployment
job creation will stop altogether. In that case, the transition will
fail,
decreases,
by being too
slow.
In the second, state workers
ment
This
rate.
A{U)
-
A)£'o)
f{Uo
+
on, the
If
{l
new
instead
>
0,
then
+
f{Uo
(1
—
A)£'o)
unemployment
all
will
0,
until
is
will also
In the third case, the case
twice. This
is
we
fail,
rate
is
stable,
drawn
and we
=
0.
At
to zero.
up
to the
that point,
the private sector will never
time from being too
shall focus on,
the case actually
unemployment
this
will take place
such that /(U)
A{U) and
If
and from then
unemployment has returned
new transformed sector will break even, and
Thus, the transition
so low as not to
be instantaneous restructuring.
then restructuring
generates
it
is
state firms will be restructured,
expand
<
no matter what the unemploy-
the horizontal locus
The outcome
private sector will
point such the
the
locus.
to restructure
when
be the case
will
intercept the
want
start.
fast.
the horizontal locus cross
in figure 4. In that case, only the lower
shall
concentrate on
it
as
we
turn to com-
parative statics.
Comparative
6.2
Using figure
and
4,
statics
we can ask how changes
in the various parameters affect
both U'
s'.
Take a decrease
in a, a goal that the Polish
government has
tried to achieve
through punitive taxation of wage increases in state firms in excess of wage norms.
Such
a decrease shifts the horizontal locus
ployment. This
is
down, increasing equilibrium unem-
because, while they have no direct effect
on production
in the
1
Analytics of Transition
2
wages makes state workers
state sector, lower state sector
bark on restructuring. As they do
so,
unemployment
less
em-
reluctant to
increases.
As f{U)/U
is
a
decreasing function of U, the exit rate from unemployment decreases until equation (6.6) holds with equality.
What happens to
creation
the speed of restructuring and to the speed of new private sector
however ambiguous. Note
is
the fact that
if
unemployment
is
first
that
what happens
to
both
linked by
is
constant, the flow into unemployment, which
is
proportional to the speed of restructuring, must be equal to the flow out of uncreation.
What
thus determined by the value of f'{U'), the effect of
unem-
employment, which
happens
to
both
is
ployment on hires
becomes negative
be signed.
If
is
equal to
in figure
1,
employment
/'(.)
is
initially positive,
unemployment high enough. And f'{U') cannot
>
f'(U*)
private sector
As we saw
at f/'.
for
new
0, if
and
in general
unemployment through wages dominate
the effects of
the effects through taxes, then higher unemployment increases both the speed
of restructuring and of private sector employment creation. But,
if
<
f'{U*)
0,
then higher unemployment leads to slower restructuring and private sector em-
ployment creation. Our numerical simulations below suggest that the
the more relevant, but this
Take now an increase in
is
case
is
ultimately an empirical matter.
saved in restructuring.
It
down, increasing equilibrium unemployment. This
is
A, a higher proportion of jobs
shifts the horizontal locus
first
because a higher A decreases the probability of unemployment after restructuring.
The
exit rate
from unemployment must thus decrease
for equilibrium to
be
maintained, and the unemployment rate must therefore increase. Note however
away
that this effect goes
workers
who know
same ambiguity
private
is
we assume
condition (5.2) instead of (5.1),
if
those
they will lose their job have to be in favor of restructuring.
at
employment
Finally, consider
if
work
for the effects
creation. If /'([/*)
an increase
locus up, leading to
in
on
>
0,
the speeds of restructuring
The
and new
then both increase.
unemployment
more unemployment. This
benefits, 6.
is
This
shifts the
A{U)
both because higher unemploy-
Analytics of Trarusition
ment
benefits
22
make unemployment
less
unattractive to state sector workers, and
because higher taxes decrease the wage in the state
has on the speed of transition
first is
sector.
The effect
this increase
The
again ambiguous, this time for two reasons.
is
the sign of f'{U'). But even
if
f'{U')
>
0,
the effect
may
still
be negative
because b affects z which enters f{U): given unemployment, higher unemploy-
ment
any given unemployment
These
a
decrease
?
generally,
should the government
is
try to affect
transition
two questions. Compared
disposal
We shall
raises
it ?
This
is
the issue to which
what
to
?
turn.
transition.
too fast or too slow,
is
we now
And assuming what
we have
first
tools at the
answer
to
government's
?
take as the objective function of the government the present discounted
value of output. Thus,
unemployment
we
focus
on
efficiency,
and give no weight
to
income or
distribution.
This leaves the issue of what the government takes as given, in solving
mization problem. There
the parameters
profitability
on
is
a sense in
we have defined
so
far.
which the goveniment can
private sector job creation, the parameter a, must
lowing the government to choose a
trivial
and
unemployment
unrealistic.
benefits
and
Thus,
fireely
we
affect
its
opti-
most of
For example, the strength of the effect of
on the organization of credit markets, which
both
unem-
the speed of transition too high or too low, and
Optimal unemployment and speed of
lb answer whether the
a
the speed of transition, should the government try to
say, raises
More
hires at
rate.
results raise the next obvious question. If a decrease in
ployment, and,
7
unemployment contributions, which reduce
benefits lead to higher
in part
the government can improve. But al-
clearly
shall
depend
makes the optimization problem
assume that the government takes the
their financing, the
wage determination
process,
and
the equation for private job creation, as given. This implies that the government
23
Analytics of Transition
cakes the function
It
can do
dN/dt =
/(f/) as given, and chooses the rate of restructuring.
our model, through changes in a, or by imposing top-down
this in
vatization and restructuring, or by modifying decision rules within firms.
not specify
how
at this point,
pri-
We shall
but will just take the speed of restructuring as the
control variable.
7. 1
Let
Optimal unemployment and speed
En
denote employment in the transformed state
problem of the government
max
is
/
sector.
The
optimization
thus:
{Ex + ENy + Ny)e-''dt
(7.1)
^0
subject to:
1
= E + En +
The
first
N + U;
equarion
is
En = K^-Eo);
dN/dt = fiU)
simply the condition that people be either employed some-
where or unemployed. The second
states the relation
between employment
in
the transformed state sector and employment in the untransformed state sector.
The third is
the relarion between private job crearion and
government takes
Using the
first
unemployment
that the
as given.
two equations to eliminate both
max(l -
A) /°°[i
Jo
E and En gives:
- U{x - Xy) + N{y -
x)]e-'*d<
(7.2)
subject to
dN/dt = f{U)
The
solution
is
characterized by an optimal
unemployment
rate,
U°, and an
as-
sociated speed of private sector job creation, f{U°). This in turn implicitly defines the optimal
speed of restructuring, which must be such that the inflow
Analytics of Transition
unemployment
into
24
is
equal to private sector job creation, and
is
thus given by
= fiU°)/il-X).
s°
If initially,
Uq < U°, then a discrete amount of restructuring takes place up
U =
the point where
U°.
If
instead,
Uq > U°, no restructuring takes place
private sector job creation has reduced
transition takes place at constant
U°
is
unemployment
to U°.
From then
until
on, the
unemployment, U°.
implicitly defined by:
x-\y = nU°){{y-x)lT)
The
intuition
U
in
given
N
is
as follows:
is
From the
associated with a change in output
unemployment can be associated with an
marginal increase in
U
-(1 — A)(x — \y).
the present value of a marginal increase in
increase in
unemployment
is
rate
A'^
if
Ay >
x,
If
the
an increase
increase in output. In addition, a
affects private sector job creation
The optimal unemployment
(7.3)
objective fijnction (7.2), a marginal increase
gains in productivity firom restructuring are large enough,
in
to
given
by f'{U°). From
(7.2),
U is given by {l-X){y-x)/r.
must be such that the net gain of a marginal
equal to zero. Dividing both sides by (1
-
A) gives
(7.3).
U°
is
characterized graphically in figure
5,
dN/dt = f{U). The oprimality condirion is
tor(x- Aj/)/(y-i).Note
is
that
if
Ay >
i,
which, once again draws the locus
that the slope of the tangent be equal
then the oprimal speed of restructuring
such as to push unemployment beyond the point which maximizes new private
sector growth,
7.2
somewhere
to the right of
A.
Comparing U' and U°
To compare actual and oprimal U's and associated speeds of
venient to rewrite the equilibrium condition
for
U'
transition,
it is
con-
in terms not of the underlying
25
Analytics of Transitian
parameters, but in terms of
tor.
Using
we
and w, the wages
and the equilibrium condition
(6.4), (6.3), (6.2)
WE-b = X{w-b) +
The equation
benefits
and costs
[f{U')/U']{w - WE)/r
as perceived
two opposite
is
not
U' — U°
1
/'(.),
the marginal effect of unemployment
w - we
which
is
is
in
that, if the decision to restructure
do not oppose
term on the
-
smaller than the difference in productivity y
optimal, leading to too low unemployment.
first
(7.4)
shows
wages between the private sector and the state
likely to hold, the incentive to start restructuring will
their jobs
and
(7.3)
sets of effects at work.
sector
is
on
unemployment f{U)/U.
we cannot. Comparing
In general,
To the extent that the difference
way
(7.4)
by workers. Note in particular that from the point
private job creation, but rather the exit rate from
sign
(5.1) yields:
has the same structure as that for U°, but this time with the private
of the workers, what matters
Can we
and the private sec-
in the state
right
it,
be weaker than
is
A second effect which goes the same
is
such that even those who
so that the relevant condition
hand side of (7.4)
x, a condition
V^ =
is
Vj/,
will lose
then the
disappears. This leads workers to be
reluctant to restructure and thus leads to a lower
unemployment
more
rate.
But two other effects work in the opposite direction: unemployment benefits make
unemployment
less painful,
leading to higher equilibrium unemployment. And,
while the marginal effect of unemployment on private job creation, f'{U)\s what
is
relevant for optimal unemployment, what matters to workers
effect
f{U)/U. By concavity of
/(.), the marginal effect
the average effect; this factor leads again to too high
Assuming we can sign
is
V — U°, can we then
sign s*
is
which
will
the average
always smaller than
unemployment
^.
— s° Here again,
t
ambiguous, and depends on the sign o( f'{U') and f'{U°).
positive,
is
If
the answer
both of them are
be the case as long as the two unemployment rates are
less
26
Analytics of Transition
than
A in figure 5,
too low an
unemployment
rate will also imply too low a speed
of transition.
These ambiguities capture the complexity of effects
biguities,
to
statements that transition
be too bold. But
we do not want
to
is
at work. In light of these
obviously too
end on
fast
this note.
am-
or too slow appear to us
We think that,
for plausible
parameter values, many of the ambiguities can be removed, and that equilibrium
speed
8
is
indeed
likely to
be too low. This
is
what we turn
to in the next section.
Cautious calibrations.
Choosing numbers
much
the various parameters of the model
for
empirical work
is
currently going on, our knowledge
where we can confidently choose parameters
fine
benchmark parameters
^°.
is
is
not easy. While
not at the point
This caveat being made, we de-
as follows:
Choosing parameters
8. 1
We
normalize the marginal product in the state sector, i to be equal to
suming that workers appropriate the returns
a =
We
1.3.
take
We
we
take the marginal product in the private sector, y to be equal to
unemployment
in the state sector.
horizoris, to
to capital in the state sector,
1.
We
benefits, 6 to be equal to
.5,
or about a third of the
take the interest rate, which here
be relatively high,
.
I
is
Astake
1.8.
wage
an index of effective
per year.
This leaves three parameters where even more guess work
is
involved.
We
take
A the proportion of state employment which can be saved by restructuring to be
equal to
.4.
This
the worst firms
is
lower than the numbers observed so
is still
to
on the output of a state
be
2,
far,
come. This implies that the direct
firm
is
to reduce
it
from
1.0 to 1.8
x
but restructuring of
effect of restructuring
.4
=
.72.
We
take c to
leading to a difference between the value of being employed in the private
.
:
27
Analytics of Transition
sector and the value of being
unemployed of
and without shame, we choose
.
1
x 2
= 20% on annual basis.
on
the effect of profit
a,
Finally,
private job creation, in
such a way as to generate roughly the speed of new private sector creation
Poland, or about
4% of the
.
I
Optimal and equilibrium unemployment and speed.
8.2
The
assumptions are given in table
results of these
ployment
condition
rates, (/*, the equilibrium
Vg = AV}v +
Ve = Vu, and U°
of
labor force per year. This yields a value of a of
in
new
,
(1
1.
unemployment
- \)Vu, U",
Table
gives three
rate obtained
unem-
assuming the
the rate obtained assuming instead
the optimal rate. For each of the three,
private job creation, f{U),
1
and the marginal
it
also gives the speed
effect of
unemployment on
that speed, f'{U).
The
first line
gives the implications of the
unemployment U*
is
benchmark parameters. Equilibrium
equal to 12%, and the rate of private sector creation
is
3.9%
of the labor force per year. This implies a total increase in private employment,
firom
both restructuring and new private sector job creation, of 6.5% per
When
restructuring decisions have to be such that those
not worse
off,
equilibrium unemployment [/**
job creation to 3%. Both
20%, which
the
/'(.)
And all
than the optimal rate of
much lower
three are
becomes negative, which
three lines give the effects of changing
unemployment
ingly,
which
rate at
lose their job are
equal to 8%, and private sector
rates are lower
associated with a speed of 4.8%.
unemployment
The next
ing
is
unemployment
is
who
unemployment
benefits lead to higher equilibrium
is
than
equal to 31%.
benefits. Increas-
unemployment.
Interest-
higher benefits and higher unemployment both imply higher taxes, and from
equation
(6.2)
and
(6.1),
lower values of being employed in either sector, and of
being unemployed. But this has nearly no effect on the speed of transition.
effects of higher
benefits
same
year.
unemployment on wages
are offset by the effects of increased
on equilibrium wages and of higher unemployment
time, higher
unemployment
The
contributions.
benefits lead to lower optimal
At the
unemployment
Table
Unemployment and Speed
1.
of Transition
U'
fiU')
U"
f{U")
U°
/(f>'°)
Benchmark
12
3.9
8
3.0
20
4.8
b
=
.6
15
3.8
10
3.1
18
4.1
b
=
.4
10
3.8
7
3.0
23
5.7
b
=
.3
9
3.6
7
3.0
26
6.7
Q =
1.4
8
3.0
5
2.1
20
4.8
Q =
1.2
20
4.8
13
4.0
20
4.8
Q =
1.1
-
-
22
5.0
20
4.8
All
numbers
eters are X
=
as percentages of the labor force.
1,
2/
=
1.8,
a =
1.3, 6
=
.5,
A
=
Benchmark values
.4,
c
=
2, r
=
.1
for the
param-
Analytics of Tramition
28
higher benefits decrease the marginal effect of a given level of unemployment
on
job creation
The
last
'*.
three lines
show how the
effects of changirig a,
one of the potential
in-
struments at the government's disposal to affect the equilibrium speed of transition.
Decreasing
a horn
1.3 in the
benchmark
and an equilibrium unemployment
studied in figure 4
earlier.
a
more unemployment,
and speed nearly equal
rate
ones. Interestingly, ftjrther decreasing
to 1.2 lead to
to I.I leads to the
second configuration
Staying in state jobs becomes so unattractive that
state firms are instantaneously restructured, leading to high
high level of unemployment, /(.)
this
is still
private sector job creation eventually reduces
9
We
to the optimal
all
unemployment. At
positive however, so that positive
unemployment
to zero.
Conclusion
have presented a model of transition where unemployment
rate of private job creation
where both
in turn affect
and the speed of restructuring of the
unemployment over
affects
both the
state sector,
and
time.
We have characterized the equilibrium rate of unemployment and speed of transition.
We
have concluded
that, while
librium speed of transition
ernment
likely to
to accelerate the transition.
wages in the
on
is
ambiguous on theoretical grounds, the equi-
state sector,
be too slow, justifying measures by the gov-
These measures may include
and top-down privatization and
firms rather than chosen by them. But
going too
fast,
and of sticking
we have
to inflexible policies,
restraints
restructuring,
on
imposed
also pointed to the dangers of
which do not reduce the speed
of restructuring in the face of high unemployment. Going too fast
may
lead to too
high unemployment and derail the transition. Inflexible policies open the scope
for self ftjlfiUing
We
have
expectations of an unsuccessful transition.
at various points in the
tential extensions. In
ending
paper discussed formalization choices, and po-
this paper,
we put
the
list
together:
29
Analytics of Transition
We
(1)
have assumed throughout that state firms were subject to a hard budget
constraint.
far in
With the exception of
bad shape,
firms in very
this
has been true so
the major Central European countries. But this should not be taken as an
assumption.
It is
will
be played
(2)
We
better thought of instead as the
for
many
have assumed,
outcome of a game, a game which
years to come.
in parallel, that the
government ran a balanced budget.
Transition has been associated with a large loss of revenues, and the temptation
have recourse to
to
inflation finance
is
So
there.
far,
the temptation has been
mostly avoided. But the same caveat applies.
(3)
We have
assumed simple interactions between the tax burden and the process
of transition, namely that the tax burden depended only
ployment. But in
fact,
on the
level of
unem-
the composition of employment and the financial posirion
of firms also matter. Profit taxes from state firms have largely vanished ^^ Collecting taxes
as
on
private firms has proven difficult.
VAT's, are not yet in place
(4)
We
have only skirted the
^*
issues of decision-making within firms.
We
taxes,
such
^^.
roles of ministries, workers, uruons,
vary across countries
New, across the board
have
restructuring and privarization.
and managers are
also skirted a
Some
firms
crucial to the
number
The
specific
outcome and
of issues by associating
have been able to partly restructure
without privatization, and some privatized firms, in particular those privarized
through buy-outs, have not restructured
(5)
We have assumed
state firms
ity.
This
is
state firms.
We
except a number of firms affected by the
that,
were identical and could survive,
not true. There
And
is
albeit at a
initial
shock,
low level of productiv-
a distribution of marginal revenue products across
the degree to which state firms can really wait and see, while
maintaining employment,
(6)
^^.
is
unclear.
have taken a black box approach
to private job creation.
Looking within
the black box, by exploring in particular the role of credit markets, could poten-
30
Analytics of TTansition
dally modify
some of our conclusions.
We also have not distinguished between two
types of private job creation, the initial burst of job creation in trade
which aimed
tor
at filling holes left
growth aimed
at
the success of the second.
services,
by the previous economic regime, and private sec-
competing with and replacing state
has been dominated by the
and
first;
much of the
firms.
The
process so far
success of the transition depends
on
1
Analytics of Transition
3
Notes
MIT
"Oxford University and EBRD, and
paper owed a
this
We
Schaffer.
thank Fabrizio
An
respectively.
early draft of
with Stanislaw Gomulka, Richard Layard and
lot to discussions
also
NBER
and
Coricelli, Peter
Diamond, Stanley
Fischer;
Mark
John Flem-
ming, Michael Gavin, Martin Hellwig, Michael Kremei; Marcus Miller, Andrei Shleifer
and Jean Tirole
'Our paper
comments and
for
is
suggestions.
short on facts, which have been
documented elsewhere. The reader
referred in particular to the papers presented at the
Eastern Europe.
pened
in
Our
analysis
Poland over the
last
is
very
much
NBER Conference on
is
Transition in
based on the interpretation of what has hap-
three years, presented in Blanchard and Dabrowski [1993],
and by the evidence on the evolution of labor markets summarized
in
Blanchard et
al.
[1993 J. Other more specific references are given along the way.
^In Poland, the
first
shock was
early 1990, the
felt in
second
in early 1991. In those
countries which started reform in early 1991, such as Czechoslovakia, the two were
roughly coincident. In most countries also, while the decline in output was sudden, the
effect
^
on employment was much slower. These additional dynamics are ignored here.
This
is
however only an interpretation,
slightly different formulation: If all finance
from employing one worker; earnings
of earnings are reinvested, and
will
be given by
as in the text.
''Chadha
{dN/dt)/N
if
will
of the private sector
*The evidence
is
is
— w — z) is the net cash flow
N{y — w — z). Thus, if a proportion
internal, ii(y
be equal to
job creation
is
proportional to investment, job creation
= a{y—w—z), arateof change rather than a change relation
As discussed below, our
et al. [1992]
is
as a retained earnings constraint leads to a
specification turns out to be
develop a model, which
is
more
a cousin of ours, but
tractable.
where the growth
based explicitly on learning by doing.
that things are
more complex.
First,
there
is
substantial turnover in
the private sectot Second, a substantial fraction of hires from the private sector
from the state sector. See Blanchard et
al.
general at the cost of analytical simplicity.
incorporated
is
is
directly
[1993]. Both aspects can be integrated, but in
One
extension which can be straightforwardly
that the pool of potential hirees includes iwt only the
unemployed but
32
Analytics of Transition
also a proportion of those
^Another
and
its
interesring extension
details of the
the private or the state sector.
in either
would be
we
interaction with the dynamics
^The
tions,
employed,
to consider the possibility of deficit finance,
analyze below.
end process depend on the comer conditions. Under our assump-
when unemployment
reaches A, the private sector closes, so that unemployment
increases discretely at that point. Thereafter the private sector remains closed, and un-
employment
*
While
ment
is
increases until everybody
B*
is
a saddle point, point
smaller than Ui, then, even
unemployed.
is
E can be
if
E
either a sink or a node.
a sink, the
is
If initial
unemploy-
economy can only be on one of
the
two paths we describe.
^This effect also plays an important role
^°Many of
in
Gavin [1993].
these countries have long collected and are
still
collecting detailed data
on
wages, flows and stocks. Results from country studies using those data are rep>orted in
Blanchard et
al.
[1993].
A clear warning
^^
ment
is
needed
at this point.
As
in the
Harris-Todaro model, unemploy-
benefits appear to be unambiguously bad in our model.
Harris-Todaro like assumption that, even
ployment
is
if
But
this
comes from our
off,
unem-
realistic
model,
the unemployed are ex-post worse
the result of the voluntary decision to restructure. In a
more
some of the unemployed would be involuntary unemployed, and the case
ment
for
unemploy-
benefits reappears.
^^A warning
that this might indeed
happen was sounded by McKinnon [1991].
^^See for example Lipton and de Combrugghe [1992] on the effects of transition and
the government budget.
^^These issues are closely related to
tion plans.
15
See
how
to
buy
off stakeholders in
top-down
privatiza-
See Shleifer and Vishny [1992].
for
example Pinto et
al.
[1993] for recent survey evidence on Fblish state firms.
33
Analytics of TTansition
References
Blanchard,
Q, Commander;
S.,
and
Coricelli, E, 1993,
Labor markets
in
Eastern Europe,
mimeo, World Bank.
Blanchard,
Q
and Dabrowski, M., 1993,
Poland, mimeo, MIT, forthcoming
MIT
Chadha,
in
Restructuring in Central Europe; the case of
"Reform
in Eastern
Europe; Pain and Progress",
Press.
B., Coricelli, E,
and growth
and Kranjak, C, 1992, Economic restructuring, unemployment
in a transition
economy, mimeo, IMF.
Gavin, M., 1993, Unemployment and the economics of gradualist policy reform, mimeo,
Coumbia
University, January.
Lipton, D. and de
Combrugghe,
A., 1992,
The government budget and
the economic
transformation of Poland, mimeo, November.
McKinnon,
R., 1991,
Pinto, B., Belka, M.,
The order of economic
and Krajewski,
S.,
liberalization,
Johns Hopkins, Baltimore.
1993, Transforming state emterprises in Poland:
Microeconomic evidence on adjustment, mimeo World Bank.
Shleifec A. and Vishny, R., 1992, Privatization in Russia
75 79
02
9
:
first steps,
mimeo.
Date Due
MOifctTta^
im
em
mi'i
4'
.^f^^^'"''
JUM.
MOV.
07
19i4
Bimi
l^^og^iS
,
Mi. 3
MAY
24
'
:'^^^
«.
mttM'^^
Lib-26-67
MIT LIBRARIES
3
TOflO
DDfl32TDS
fi