Disorganization

Digitized

by

the

Internet

Archive

in

2011

with

funding

from

Boston

Library

Consortium

Member

Libraries

|/io.%-30

®

working

paper

department

of

economics

DISORGANIZATION

96-30

OlivierBlanchard + Michael KMWvcf"

October, 1996

massachusetts

institute

of

technology

50 memorial

drive

Cambridge, mass.

02139

OlivierBlanchard » ^ctaxel Ktev*t

Olivier

Blanchard

Michael

Kremer

*

* _{Department of Economics, MIT, Cambridge,}

MA

_{02138. (blanchar@mit.edu) and}

(kre-mer@mit.edu).

We

thank SusanAthey, Abhijit Banerjee,SimonCommander,BillEasterly,Avner

Greif, Dani Kaufmann,Gian-Maria Milesi-Ferreti,Wolfgang Pesendorfer, Julio Rotemberg and

Andrei Shleifer for discussions and comments.

We

thank Maciej Dudek and Andrei Sarychev

for researchassistance. Michael Kremerthanks the Research Departmentof the International

Disorganization

Abstract

Undercentral planning,

many

firms relied

on

a single supplier for

many

oftheir inputs. Transitionhasled to decentralizedbargaining betweensuppliers and

buy-ers.

Under

incompletecontracts orasymmetric information,suchbargainingcan

beinefficient, andthemechanismsthatmitigatethisproblemintheWestcanonly

play a limitedrolein transition. Forexample, thescope forlong-termrelations to

reduce opportunistic behavior is limited

when many

existing firms are expected

to disappear orchange most oftheir suppliers inthe future.

The

result hasbeen

disorganization, anda sharp decrease inoutput.

The

empirical evidencesuggests

thatdisorganization has played amore important role in theformer Soviet

Union

Figures 1

and

2

show

the evolution ofofficial measures of

GDP

in the countries

ofthe former Soviet Union, since 1989.

The

two

figuresgive a striking picture,

that of

an

extremely largedecline inoutput. In 10 out of the 15 countries,

GDP

for 1996is expected tostand at less thanhalfits 1989 level.

The

evolution of theseofficial measuresreflectsin large part ashift to unofficial,

unreported activities. But studies

which

attempt to adjust for unofficial activity

stillconclude thatthere has

been

a largedecline inoutput.Forexample,arecent

study, by

Kaufmann

and

Kaliberda [1995], puts the actual decline in

GDP

from

1989to

1994

forthesetofcountries of theformerSoviet

Union

at

35%, compared

toa reported decline ofabout

50%.

x

Thislarge declinein outputpresents

an

obvious challenge forneo-classical

the-ory.

Given

themyriadprice

and

trade distortionspresent undertheSoviet system,

these countries surely operated far inside theirproduction frontier before

tran-sition.

With

the removal of these distortions, one

would have

expected

them

to

operatecloser to the frontier,

and

thusoutput toincrease.

And

one

would have

expectedthat the

development

of

new

activities

would

shift theproduction

fron-tierout,leading to further increases inoutput over time.Thisclearlyisnot

what

has happened.

In thispaper,

we

explore

one

explanationforthe decline of output,

which

we

be-lievehas played

an

importantrole.

The

best

word

todescribeit isdisorganization.

1. Estimatesforanumberofotherstudies aregivenanddiscussedin European Bankfor

CO CJ

o

m

cd

a

CD

o

L_ ZD CO 3) L_

a

CD

m

CO CO =)

Q

CJ

o

cz

o

o

>

UJ CD CD [_ ZJ

en

as CD CD LD CD CO CD CD CD CO -r co

O

CD

—

o

CD CD

O

o

CO

o

CO

O

cq

o

co LO

o

CO

o

o

o

CM CO

o

a

CD

>-|8ASH

GO i_ CD

O

CD 00 CJ ZJ

Q_

CD

—

^

a, CD CD

n

cd Q_

Q

CD

O

cz CD C\J CD C\J I I I

mCKM(ESj-)UJO

m3<»-o;S<muj

LD O, 00 oo CD CD CD CO CD

O

CD CO CD CM CO CO CD CD CD

O

CD CD

a

CD >-|9A9-|

Disorganization

Itslogic goesas follows:

• Central planning

was

characterized by a

complex

set of highly specific

re-lations

between

firms. (Following standard usage,

we

shall call a relation

"specific" if there is a joint surplus to the parties from dealing with each

other rather than taking their next best alternative.)

There

were typically

fewer firms in

each

industry thanin the West.2 For

many

inputs, firms

had

or

knew

ofonly

one

supplier from

which

to buy. For

many

oftheir goods,

firms

had

or

knew

ofonly

one

buyer to

whom

tosell.3

• Specificity opens

room

forbargaining.

Under

asymmetric informationor

in-complete contracts, the result of bargaining

may

be inefficient.

A

complex

set of highly specific relations can thus lead to a poor aggegate outcome,

with large effects

on

total production.

•

Through

variousinstitutions

and

arrangements,

economies

develop waysof

limiting the adverse effectsofspecificity:

Under

central planning, the

main

instrumentwas the coercive

power

ofthe

central planner toenforce production

and

delivery ofgoods.

In theWest, for

some

goods, there are

many

alternative buyers

and

sellers,

eliminating specificity altogether.

And

for other goods, the scope for ad-verseeffectsofspecificityis reduced by arrangementsranging fromvertical

integration, to contracts, tolong-term relations

between

firms.

• Transition eliminated the central planner,

and

thus the

main

instrument

2.

One

reason is that, in the West, (non natural) monopolies and associated monopoly rents

createincentivesforentry ofnewfirms.There was no suchmechanismundercentralplanning.

3. The buyerwas often acentralized trade organization, which disappeared altogether during

used to limit the adverse effects ofspecificity under the previous regime.

But the

mechanisms which

exist in the

West

could not operateright away.

New

buyers

and

sellerscannot becreated overnight.

The

roleofcontractsis

reduced

when

many

firmsare close tobankruptcy.

At

the

same

time,

tran-sition has created the anticipation that

many

state firms will disappear or

change

suppliers, thus shortening horizons

and

thescope for long-term

re-lations to avoid the adverse effectsofspecificity.

•

The

result has

been

the

breakdown

of

many

economic

relations. Trade

between

the republics of the former Soviet

Union

has fallen by far

more

than

seems

consistent with efficient reallocation. Despite price

liberaliza-tion,

many

firms report shortages of inputs

and

raw materials. Firms lose

crucial workers/managers,

who may

have

been

their only

hope

for

restruc-turing

and

survival. Cannibalization of

machines

iswidespread,

even

when

it appears that

machines

would

be

more

productive in their original use.

Thisgeneral disorganization has played

an

important role inexplaining the decline inoutput.

Our

paperis organizedas follows:

• In sections 1 to 3,

we show

throughthreeexamples

how

specificity, together

witheitherincompletenessof contractsorasymmetricinformation,

can

lead

to large outputlosses.

We

show

how

theeffects

depend

both

on

the degree

ofspecificity in the relations

between

firms,

and

on

the complexity of the

production process.

We

show

also

how

the

emergence

of

new

private

op-portunities

can

lead to a collapse ofproductionin the statesector,

and

toa

sharpreductionin total output.

Disorganization

specificity take time to develop

and

have therefore playeda limited role in

transition.

We

also discuss the policy implications ofour theory,

and

argue

thatit

makes

a

—

limited

—

case forgradualism.

• Insection 5, taking measures ofreported "shortages" of inputs by firms as

indicatorsofdisorganization,

we show

thatdisorganization appearsto

have

played

an

important role in the former Soviet Union, but a

more

limited

one

inCentral Europe.

Before starting, it

may

be useful to

do

some

productdifferentiation.

Our

analysisisbased

on

standardideas inmicroeconomics, from asymmetric

in-formation, to hold-up problems under incomplete contracts, to the

breakdown

ofcooperation in repeated

games

as the horizon shortens or as alternative

op-portunities arise.4

The

potential importance ofspecificity for

macroeconomics

has

been

emphasizedrecentlybyCaballero

and

Hammour

[1996].

Our

contribu-tion is to point to the potential importance of these problems in the context of

transition.

Our

analysis

complements

other explanations of the decline inoutputin

transi-tion.

Most

explanationsstart from the

need

for state firms to decline

and

trans-form,

and

for

new

private firms to emerge.

Some

argue that the output decline

may

reflect efficient reorganization

and unmeasured

investment in information

capital, asfirmstry

new

techniques

and

workers lookforthe rightjobs. (Atkeson

and Kehoe

[1992],

Shimer

[1995] forexample). Incontrast,

we

seethe evidence

4.

On

this last topic, see for example Wells and Maher [1995] for breakdowns of

coopera-tion within households, and Banerjee and

Newman

[1993] for an example in the contextof

as reflecting in large partinefficientdisorganization.

Other

explanations

combine

reallocationwith

some

remainingdistortion, explaining the declinein output as a second-best outcome. Perhaps the

most

widely accepted explanation argues

that efficient reallocation

would have

implied a decrease inreal wages of

some

workersinconsistent witheither actual

wage

determinationor politicalconcerns

about

income

distribution;asaresultof thesecontraints,

employment and

output

have decreased (see Blanchard [1996]).

We

read the evidence from the former

Soviet

Union

as

showing

that thereis

more

atwork,

namely

disorganization.

Two

theories are related to ours.

The

first, developed by

Murphy

et al. [1992],

em-phasizesthe adverseeffectsofpartialprice liberalization.

The

second, developed

byShleifer

and

Vishny [1993], focuses

on

the roleof corruption,

and

emphasizes

the adverse effectsof rent extractionby

government

officials.

We

shall point out

the relations as

we

go along.

1

A

chain

of

production

This

and

thenext

two

sectionsdevelopthreeexamples,all three

aimed

at

captur-ing different

mechanisms

through

which

specificities

can

lead toa large decline

inoutput during transition.

The

first is also the simplest. Itgoes as follows:

•

A

good

is

produced

accordingto aLeontieftechnology

and

requires

n

steps

of production.

Each

step of production is carried out by a different state firm

(A

point ofsemantics here.

We

shall use "state" firms to denote the

firms

which

existed under central planning

and produced

under the plan,

whether

ornot they have

been

privatizedsince the beginning of transition.

Disorganization

since transition.)

One

unit of the primary

good

leads, after the

n

steps, to

one

unit of the final good. Intermediate goods

have

a value ofzero.

The

price of thefinal

good

is normalized toone.

One

may

replace "production" by "distribution" here:

an

interpretation

with a Soviet

Union

flavor is that ofa

good

having to travel through

n

provinces/republics,togo from theprimaryproducertothefinaldistributor.

•

The

supplier ofthe primary

good

has

an

alternative use for the input equal

to c.

One

can

think ofc as a privateopportunity, perhaps the production of a

much

simplergood, avoiding the divisionof laborimplicit in thechainof production; c

may

therefore be

much

lower than1. (It isstraightforward to

introducesimilar privateopportunitiesfortheintermediateproducersalong

thechainofproduction. Butthe point is

made

more

simplyby ignoringthis

possibility.)

•

Each

buyer alongthechain

knows

only thesupplierit waspairedwithunder

centralplanning,

and

vice-versa.

The

end

ofcentralplanningthus leaves

n

bargainingproblems.

We

assume

that thereis

Nash

bargaining at

each

step

ofproduction, withequal division ofthe surplusfrom the match.

The

characterizationofthe solutionisstraightforward

and

isobtained by working

backward

from the laststage ofproduction.

The

value of the surplus in the last

bargainingproblem,

between

thefinalproducer

and

thenext-to-lastintermediate

producer isequal to 1 (by assumption, the

good

is useless before thefinal stage)

.

Thus, the next-to-last intermediate producerreceives (1/2). Solving recursively,

the first intermediateproducer receives(l/2)n.

The

surplus to be divided

between

the first intermediate producer

and

the

positive

and

production takes place along the chain. Ifinstead c

>

(1/2)", the

primary producerprefers totake

up

hisprivateopportunity.Thus, theappearance

of

even

mediocreprivate opportunitieswill lead to thecollapse ofproductionin

thestatesector: the decreasein total output

may

beas large as 1

—

(l/2)n inour

example

(what matters here is not the absolute level of private opportunities,

but their level relative to opportunities in the state sector.)

The more

complex

the structure ofproduction (thehigher n), the smaller the private opportunities

needed

totrigger the collapse ofthestate sector.

Avoiding the adverseeffectsofspecificityin

complex

structuresofproductionis

a problem faced by firms in the

West

as well.

Even

in the West,

many

goods are

sufficientlyspecificthat firms

can

onlyget

them

from

one

supplier, orsell

them

to

one

buyer.

We

shall discuss in Section 4 below the

mechanisms which

alleviate

the adverseeffectsofspecificity in the West,

and

why

they

have

played alimited

role in transition. Buta few points are worth

making

here already.

The

source of the inefficiency

and

ofthe collapse ofoutput in our

example

is

a hold-up problem: each intermediate producer

must

produce its intermediate

good

before bargaining with the next producer along the chain.

Once

he has

produced

the intermediate

good

and

has

no

alternative use forit, hisreservation

value isequal to zero.

The

inefficiency

would

therefore disappear ifall theproducers could sign

an

en-forceable contract before production took place. In thiscase, production

would

take place so long as cwas less than 1,

and

the producers as a

whole

would

re-ceive 1

—

cof the output.

The

source of the problem is thus the combinationof

Disorganization

8

The

assumptionofincompletecontracts

seems

quitereasonableinthe case of the

former Soviet Union.

The

legalsystemisstillinconstruction. Contract

enforce-ment

is notoriouslyweak. Firmsoften

have

fewvaluableassets,

making

it hardto

punish

them

for deviating fromthe contract.

Another

natural solutionisvertical integration. Inourexample,

even

partial

in-tegration (which effectively reduces the

number

ofbargaining problems

down

from

n

to a smaller

number)

, alleviates the problem,

and

full vertical

integra-tion (where one decision-maker decides

whether

or not to produce) eliminates

it.

Even

short ofvertical integration, long-termrelations

between

producers

can

also alleviate the problem. Iftheyare

engaged

ina long-term relation, producers

may

be willing topay their supplier

more

than the price

we

derived above,

mak-ing it

more

likely that production takes place in the chain.

As we

shall argue in

Section 4 however, the nature oftransition issuch that it reduces thescope for

vertical integration

and

forlong-term relations toinduce suchbehavior.

2

A

state

firm

and

its

suppliers.

I

Inour second example, thesource ofinefficiency isnot incompletecontractsbut

asymmetric information:

•

A

state firm needs

n

inputs in order to produce. Ifall inputs are available,

the firm

can

produce

n

units ofoutput. Otherwise, output isequal tozero.

•

Each

input issuppliedby

one

supplier.

Each

supplier has

an

alternative use

for his input, with value c, distributed uniformly

on

[0,c]. Let F(.) denote

the distribution function, so that

F(0)

=

and

F(c)

=

1.

Draws

are

We

can

again think ofc as a private sector alternative, such assmall scale

production or sale of the input to a foreign buyer.

The

distribution of c is

known,

but the specific realizationof

each

cis private information to

each

supplier, c

can

bethoughtofasindicating thedegree of

development

ofthe

private sector,

which

we

assume

isvery low pre-transition,

and

increasing

overtime thereafter.

•

The

statefirm maximizes expected profit. It

announces

a take-it or leave-it

price

p

to

each

supplier (given the

symmetry

built in the assumptions, the

priceisthe

same

forallsuppliers). Ifthepriceexceedsthe reservationprices

ofallsuppliers, production takes place in thestate firm. Otherwise, it does

not,

and

all suppliers use their

own

private opportunity.

We

can

now

characterizeexpectedstate, private,

and

total productionasa

func-tionofc

and

n.

We

must

firstsolve fortheprofit-maximizingpricesetbythestate

firm.

Given

price _p,expected profit isgivenby:

ir

=

(F(p))

n

(l-p)n

The

firsttermis theprobability that productiontakesplace.

The

second isequal

toprofit ifproduction takesplace. Maximizing with respect to

_p

yields theprofit

maximizingprice:

p

=

min(c

,

—

—

_-)

n

₊

I

The

firm never pays

more

than the

maximum

alternative opportunity, c. But, if

Disorganization 10

increase its price. Increasing the price

would

increase the probability that

pro-duction takesplace, but

would

decrease expectedprofit.

Given

this price, expectedstate production,

Y

s, isequal to:

n

1 n

Y

s

=

nmin(l

, (

——

-) ₎

n

+

lc

Expected

private production,

Y

_p, isequal to theprobabilitythat stateproduction

does not take place, times the conditional expected

sum

ofalternative

opportu-nities (conditional

on

at least

one

private alternative being largerthan the price

offeredby the state firm).

Y

p

can

be writtenas:

nr

n

\ n+1

^

=

^(0,1-^1)

₎

Expected

totalproduction isin turnequal to

Y

=

Y

s

+

Y

p

The

best

way

to see

what

theseequationsimplyisto lookatFigure3,

which

plots

the behavior, for

n

'.==

4, ofexpectedstate, private

and

total production (all three

normalizedby the

number

ofinputs, n) for values ofcranging from 0.4 to 2.4.

To

interpretthe figure,itisuseful tokeep theefficient

benchmark

in

mind

(which

would

bethe

outcome

ifthemonopsonistwasfullyinformedaboutthealternative

opportunities ofeach supplier

and

thus was fully discriminating, or ifthere were

Z3 CL Z3

o

~o

o

a

o

cu

o

>

a>

o

en a;

o

CU o_

X

CD

o

>

o

CO CD cu

3

en

o

O

J2> ao

O

o

fl

Z' I

OL

80

9'0 *'0 3 -0"0

Disorganization 1 1

was less than one, production

would

always take place in the state sector.

As

c

increased

above

one, it

would

sometimes beefficient not toproduce inthe state

sector:expected productioninthestatesector

would

decrease,but totalexpected

production

would

increase. Eventually, asc

became

verylarge,

most

production

would

take placeinthe privatesector, with expected total productionincreasing

linearlywith c.

In contrast with the efficient outcome, increases in private opportunities lead

initially to such a large decrease in state production that the net effect is a

decrease in total production. Total production starts declining

when

c exceeds

(n/(n

+

1))

=

.8. For cequal to 1, the efficient

outcome would

be that

produc-tionstill only takes place inthestate firm, with totalproduction equal to

n

(1 in

the figure, as output isdivided by the

number

ofinputs).

The

actual

outcome

is

aproductionlevel ofonly0.75. It isnotbefore chas reachedavalue ofabout 2.0

that total production isagainequal to its pre-transitionlevel.

Complexity

ofproduction in the statesector increases the size ofthe decline in

output.

We

can

think of

n

as

an

indexof thecomplexityoftheproductionprocess.

Figure 4

shows

the effectsof alternativevalues of

n

on

theevolution ofexpected

total production (again normalized by the

number

ofinputs) as a function ofc.

The

larger n, the higher the price that the state firm offers: for c large

enough

so that the

minimum

condition does not bind,

_p

=

n/(n

₊

1). Thus, the higher

n, the higher the value ofc required to trigger the collapse ofthe state sector.

But, also the higher n, the smaller the probability that production takes place

in the state firm for c

>

n/(n

+

1),

and

thus the greater the initial collapse of

total productionas c increases.

As

n

approachesinfinity, the priceoffered by the

O

>

CD CD "O

o

Z3

o

o

o

CD

o

CD Q.

X

CD

>

CD CD CD ZJ

O

CNI CO CD

o

o

O

LZ >'L

£1

S'L

11

0*1

60

9*0

L0

9*0

Disorganization 12

above one: the probability that at least

one

supplier will

have

a realizationofc

greater than

one

goes to one. Thus, as

n

approaches infinity, the path ofstate

production (normalizedby the

number

ofinputs) approaches

one

forc

<

1,

and

c/2 (theexpectedvalue ofprivateproduction) forc

>

1.

When

cincreasesabove

1, output fallsby fifty percent.

Puttingourresults inwords: Pre-transition, suppliers have such pooralternative

opportunities that the state firm can offer a price at

which

there is

no

better

alternative than to supply.

As

transition starts,

and

alternatives improve,

some

suppliers

now

have

more

attractive opportunities.

They

start asking the firm for

more.

Not

knowing which

suppliersare bluffing

and which

are not, thefirmhas

no

choiceother than toofferagivenprice

and

take theriskofnotgettingthe inputs.

The

result is

an

initial decrease in total production asopportunities improve.

Let us take

two

sets ofissuesat thispoint:

(1) This

example

isclosely related to the analysisby

Murphy

et al. [1992] of the

effects ofpartial priceliberalization. In theirmodel, the fact that

some

pricesare

held fixed while others are left free

can

also lead to

an

inefficient diversion of

output from state to private firms (Young [1996] argues that a similar perverse

mechanism

isrelevantinChina).

Our

argument

isthat the

same

perverseeffects

occur

even

underfullpriceliberalization.

The

presenceofreported shortages (see

section5) ina

number

offormerrepublics a

number

of yearsafternearlycomplete

price liberalizationsuggests that there is indeed

more

at

work

than partial price

liberalization.

The

implications of the

two

models for policy are also potentially

quite different.

The Murphy

et al. [1992]

model

implies the desirability offull

be lower during transitionthan it

was

under central planning.

(2)

We

have

assumed

that thesellers

had

privateinformation,

and

that theseller

set theprice.

What

isessentialhereisthat therebe privateinformation,

and

that

the

uninformed

partyhave

some

bargaining power.

Suppose

for

example

that the

firm has private information about the value offinal production, but that, now,

suppliers

make

price offers to the firm.

Assume

that theyhave

no

private

oppor-tunities,

and

choose the priceso as to

maximize

their expected revenue. Ifthe

sum

of the prices charged by suppliers is higher than the actual value of final

production, production does not take place. It is straightforward to

show

that,

in thiscase, the probability that suppliers collectivelyask for too

much

and

that

final production does not take place is positive. It isalso easy to

show

that this

probability increases in the

number

ofsuppliers: the

more complex

the

produc-tionprocess, the

more

likely isstate production to collapse,

even

in the absence

ofprivate opportunities.5 This result is closely related to results

on

the adverse

effectsofcorruption

on

output.

We

have

assumed

thattheprice setterswere

sup-pliers. But,absent theruleoflaw,they

may

be

government

officialssellingpermits

tooperate thefirm.

There

istherefore a close parallel

between

our

argument

that

decentralizedbargaining

between

firmshasled to adecline inoutput,

and

the

ar-gument

made

by Shleifer

and

Vishny [1993] that political decentralization has

5. More formally, assumethatifallinputs are supplied, outputisequal tone,withedistributed

forexampleuniformlyon [0, 1].

Now

consider theproblemfacing supplieri. Denote thesumof

theprices setbyallothersuppliers asp_,-.Supplierimaximizes(1

—

(p_,-

+

Vi)ln)Pi< ar|d

^

us

chooses apriceequal to pi

=

(n

—

p_,)/2. In thesymmetric equilibriump_,-

=

(n

—

l)p,-,so

that the

common

price isequalto_p

=

n/(n

+

1),and thestate firm produceswith probability

Disorganization 14

led tocompetition

among

rent extractors, withlarge adverseeffects

on

output.6

3

A

state

firm

and

its

suppliers.

11

The

third

example shows

how

specificity

and

complexity

can

combine

tocreate

coordinationfailures,

and

a collapse ofoutput.

In this model, technology is again Leontiefin

n

inputs. Because the

interpreta-tion is

more

natural here,

we

think of the

n

suppliers as the

n

crucial workers in

the firm. But this is not essential.

The

essential difference

between

this

and

the previous

model

is inthe timing ofdecisions. In the previous model, ifstate

pro-ductiondidnot takeplace,supplierscouldstilltake uptheir private opportunity.

Here, suppliers

have

to decide

whether

or not to take their private opportunity

before they

know

whether

production is taking place in thestate firm. This

dif-ference in timing introduces problems ofcoordination

between

workers,

which

we

now

examine.

The

model

isas follows:

•

A

statefirmneeds

n

workersinorder toproduceefficiently.

Assume

that its

technologyisLeontiefin thoseworkers

(who

aretherefore theworkers

who

are difficult to replace;

we

can think of the other workers as having

been

solved out of the production function)

.

Ifallworkersstay,thefirmproduces

n

unitsof output, equivalently 1 unitof

output per worker. If

one

or

more

workers leave, the firm can hire

replace-6. Empiricalevidenceon thenumberand thesizeofbribesneeded torunabusinessinUkraine

merits; replacementsare

however

lessproductive. Ifthefirm hastohire

one

or

more

replacements,outputper workerisequal to

₇

<

1.

Thus

n

measures the complexity ofthe productionprocess.

The

parameter

7

(whichwasimplicitly put equal to zero inthe first example) is

an

inverse

measure

of the specificityofexistingworkers.

•

Each

workerhas

an

alternativeopportunitygiven byc,distributeduniformly

on

[0, c].

Draws

are independent across workers.

The

distribution of c is

known,

but the specific realizationof

each

cisprivate information.

We

can

thinkofcasrepresentingtheopportunities

open

tothe

most

qualified

work-ersafter transition, fromopeningtheir

own

business, toworking forforeign

consultingfirms,

and

so on.7

•

_The

wage

paid is equal for all workers (this follows from the assumption

that alternativeopportunitiesare private information,

and

the

symmetry

in

production),

and

isequal to output per worker. This assumptionsimplifies

the analysis,

and

is reasonable in the case of

many

former state firms, but

the analysis

would

be qualitatively similar as long as wages increased with

other workers' participation.

•

Workers must

decide whether to take

up

their private opportunity before

they

know

thedecisions ofother workers,

and

thus the levelofoutput

and

the

wage

perworkerin thestate firm.

We

assume that theyare riskneutral,

sothat their decision isbased

on

the expected wage.

7.

An

alternative interpretationofthismodelisasamodelofshirking, where workers have

dif-ferent costsofeffort.If allworkersputeffort,outputperworkerisone.Ifsomedonot,outputper

workerisequal to₇instead. Underthat interpretation,onemayinterpret themodelasapplying

Disorganization 16

Thus, if all workers decide to stay, output per worker,

and

therefore the

wage, isequal to one. If

one

or

more

workers take

up

their private

alterna-tive,

some

replacement workers

must

be hired,

and

output per worker

and

the

wage

are equal to 7.

It isclear that the solution to the problem facing each worker has the following form: Stay if c is less than

some

threshold c*, otherwise leave.

To

characterize

c*, think of the decision problem facing one worker. Ifhe leaves

and

takes

up

hisalternativeopportunity, he will receive c.

Suppose

hedecides instead tostay.

Given

that, by symmetry, the (n

—

1) other workers also

have

a thresholdequal

to c*, the probability that they all stay is given by (F(c*))

n_1

.

Expected

output

per worker,

and

thus the

wage

he

can

expect to receive ifhe stays, is thus equal

to (F(c*))n_1

+

(1

-

(F(c*))

n-1

)7.

The

threshold level c* issuch that he

must

be indifferent

between

leavingor staying,so that:

» n—

1

c*

=

₇

+

min(l,(-)

)(l-

₇

) (3-1)

c

Expected

output per worker,

and

thus the expected wage, are givenin turnby:J

io

=

₇

+

min(l,

fy

)(l-7)

(3-2)

8. Notethedifferentexponentsin (3.1) and (3.2),(n

—

1)versusn.Each workerknowshisown

alternativeopportunitywhen choosingwhethertostayornot, and thus the uncertaintycomes

fromwhatthe n

—

1otherworkerswilldo.Theexpectedwage isthe unconditional expectation

ofthewage anddepends onwhatallnworkersdo.Thedifferencebetweenc* and

w

goestozero

Disorganization 1

The

solutiontoequation (3.1)

can

be

one

of three types,

depending

on

the value

of c.

These

three cases are represented in Figure 5,

which

plots both sides of

equation (3.1)

on

thevertical axisagainstc*

on

the horizontal axis.

The

figureis

drawn

assumingvalues of

n

=

5,

and 7

=

.2.

The

left-hand-side of the equation

isrepresented bya45-degree line.

The

right

hand

side isrepresentedby acurve

which

isinitiallyconvex, turning flat at c*

=

1.

• Forlowalternativeopportunities, thereisonly

one

equilibrium,in

which

all

initial workersdecide to stay, leading to a level ofoutput per workerequal

toone.

InFigure 5, the only equilibrium corresponding tothe case

where

c

=

.3is at point A.

• For intermediate alternative opportunities, there are two equilibria (the

equilibriumin-betweenisunstable usingstandardarguments),

one

in

which

all workersstay

and

output per workerisone,

and one

in

which most

work-ersarelikely to leave,

where

c*,

and

inturnexpected outputper worker, are close to 7.

InFigure5, forcequal to .7, the twoequilibria are atpoints

A

and

B.

At A,

c* is 1.

At

B, c* (and the expected

wage

from (3.2)) isveryclose to .2: the

probability thatat least

one

workerwill leave isveryhigh.

• For higher alternative opportunities

—

for c

>

1

—

there is a unique, low activity, equilibrium.

In Figure 5, for c

=

1.1, the only equilibrium is at point B,

where

c*,

and

the expected wage, areboth equal to a valueclose to .2: production in the

statesectoris veryunlikely.

lo CD

O

>

CD

>

O

CD * CD Z3 ^_ -Q Z5 CT UJ LO CD ZS

O

oo

d

CO *.

d

<->

^3-d

CNI

d

o

o

I

01

80

90

V0

Z'O O'O

*D

(1)

The

first istheexistence, over

some

range,ofmultipleequilibria. This result

isnot

due

toprivateinformation.

To

see this, note that

even

when

workers

have

an

alternativeopportunity equal toa

known

valuec,then, forvalues ofc

between

7

and

c, thereare

two

equilibria,

one

in

which

all workers stay in the state firm (the efficient

outcome)

,

and one

in

which

all workers take

up

their alternative

opportunity.

Is

one

equilibrium

more

likely to prevail?

The

arguments here are standard.

On

the

one

hand, theequilibrium with high productionisParetosuperior.

The game

is

one

ofpurecoordination, so ifthe workers

can

talk beforehand, theyshould

be able to coordinate

on

the equilibria

which

is better for all of them.

On

the

otherhand,ifeach workerbelievesthatthere is

some

probabilitythat

some

other

worker will irrationally refuse to participate, then the high-equilibrium

can

un-ravel.

As

n becomes

large,

even

a small probability that

one

worker will

irra-tionallyrefuse to participate

can

lead to the rational collapse ofoutput.

This aspect of the

model

captures the fact that

complex

production processes,

which

require the contribution of

many

specific workers (or/and

many

specific

suppliers) imply

complex

coordination problems.

A

reflection of these problems

is the potential presence of multiple equilibria.

The

argument was

first

made

by

Bryant [1983] in the contextof

Western

Economies. But the prevalence of

spe-cific relations (and

we

shall argue in the next section, the lack of

mechanisms

designedtohandle suchcoordination problems)

makes

it

more

relevantfor

coun-triesin transition.

Some

firms

may

disappearsimplybecause theyareexpectedto,

either bytheir suppliers or their essential workers.

informa-Disorganization 19

tion

and

interactions

between

decisions given the timing ofdecisions. It is the rapidcollapse ofstate outputas

soon

asthe probability that at least

one

private

alternativeexceedsthe

wage

underefficientproductioninthestatefirm

becomes

positive.Forexample,forthecase

where

c

=

1.1inFigure5 (sothat theexpected

value of private opportunitiesisonlyequal to.55,

compared

to 1 inthestatefirm

if workers stay) the probability that all initial workers stay in the state firm is

nearly equal to zero,

and

expected productionperworkerin thestate firmisvery

close to .2, a highly inefficientoutcome.

To

see

why

output collapses, note that as

soon

ascisgreaterthanone, there isa

chance

that aworkerwill leave.

The

wage

thatremaining workers

can

expect to

receive is thus reduced in proportion to the probability thatat least one worker

will leave. Inturn, thisreduction intheexpected

wage

prompts

more

workers to

leave,

and

so on.

The

effect ofhigher values of

n

is to magnify the interactions

between

the workers' decisions. For example, for c

=

1.01, expected output per

worker inthe state firm isequal to .92 if

n—

2, butfalls to .21

when

n=3.

9

This aspect of the

model

appears to capture a

number

of aspects oftransition.

It captures the idea that crucial

managers

may

leave a firm, despite potentially

large payoffs from turning it around, because they are not sure that others will

stay

around

longenough. If

we

returnto

an

interpretationof the

model

with

sup-9. Toseetheroleoftimingdecisions, theseresultscanbecompared tothosewhichobtainwhen

workers decide whetherornotto taketheir private opportunity afterhavingobserved the

de-cisionsofothers. In that case, the probability thatallworkersstay in the state firm isequal to

(l/c)n, and expected outputperworkerin thefirmisequal to₇

+

(l/c)"(l

—

7).Forc

=

1.01,

pliers rather than workers, the

model

isconsistent with anecdotal evidence that

suppliers are often very reluctant tosupplystatefirms, because ofthedifficulties

ofgetting paid. Ifthey are not paid

on

delivery but only ifthe firm succeedsin

producing

and

sellingitsproduction, then,just like workersinourexample,

sup-pliers

become

residual claimants, (this raises the issue of

why

state firms

do

not

pay theirsuppliers

on

delivery.

More

on

thisinthenextsection.)

The

probability

that

one

of

them

will notsupplyleadsall of

them

tobe learyofsupplying,leading

tocollapse of the statefirm.

4

Managing

specificity

We

have

shown

throughthreeexamples

how

specificity

and

complexity

may

have

combined

toleadtoa largedecrease ofoutputduring transition. Butthese

exam-plesraiseageneral issue.Ifanything, theproductionstructure inthe

West

is

even

more complex

thanit isin theEast.

Why

aren'tproblems arising fromspecificity

asseverein theWest?

We

have

alreadyhintedat

some

ofthe answers:

(1) First,intheWest,

many

goodsaretradedinthick markets, markets with

many

buyers

and

many

sellers.

And

while for

many

intermediate goods, firmsdeal with

only

one

supplier, there are likely to be other potential suppliers,

and

shifting

suppliers

may

be difficultbut not impossible.

For

many

goods however, thick marketscannot be created overnight.

There

are

just not

enough

suppliers orbuyersin the

economy

tostart with;

new

firms

must

Disorganization 21

increases the potential

number

of buyers

and

sellers

can

clearly help here.

In-deed,

one

of the arguments for imposingconvertibility

and

low tariffs in Poland

in

1990

was

indeed to limit the

monopoly power

ofdomestic firmsby increasing

competition,

an argument

closelyrelated to ours (Sachs [1993]).

(2) Second,thereis

more

scopeforcontractstohandle problemscreatedby

speci-ficity.

The

potential role ofcontractsinsolving or alleviatingproblemsofspecificity is

clearest in our first example. In that model, ifproducers

can

sign binding

con-tractsbefore productiontakesplace, thentheywill avoid hold-upproblems,

and

achieve

an

efficient outcome. Total output will not decline

when

private

oppor-tunities improve.

There

is little question

however

that the scope for contracts to alleviate

prob-lemssuchas hold-upsis

more

limited inthe East.

The

deficienciesofthe current

Russian legalsystem

have been

pointedout by

many

(for

example

Greif

and

Kan-del [1994]): Contractenforcement is

weak

(in our example, producers

have an

incentive ex-post to renege

and

offer a lower price to their suppliers. Contract

enforcementistherefore crucial).Firms have few assets

which can

beseized,

and

thus less to lose from breaking a contract. This last aspect takes us to the next

point.

(3)

Most

statefirms in transitionare short

on

"cash" (moreprecisely

have

limited

liquid assets,

and

limited access tocredit.)

The

potential roleofcashinreducingproblems

coming

fromspecificityisclearest

in our third example. Consider the case

where

there are

two

equilibria,

one

in

forcing thefirm to hireless efficientreplacements.

A

commitment

by the firm to

pay

each

existingworkera

wage

ofone, independentlyof

what

theother workers

do, will lead all workers to stay,

and

will sustain the efficient equilibrium, at

no

cost to the firm.

Thus, firmswith "deeppockets"can alleviate

some

of the problems arising from

specificity.10 But state firms in the former Soviet

Union

typically

do

not

have

such deep pockets: wages typically exhaust

most

of revenues, firms have limited

access tocredit,

and

many

are close tobankruptcy.

Under

these conditions, they

have

no

easy

way

ofconvincing theirworkers tostayor their suppliers to supply.

This

argument

points to the role offoreign direct investment: foreign firms

typ-ically have

deep

pockets.

They

can paysuppliers

on

delivery; they

can

credibly

promise to pay theirworkers

even

ifthingsgo wrong.

They

can

therefore avoid

theproblemsof coordination facedbystate firms.

(4)

While

state firms were often

more

vertically integrated than in the West,

the type ofvertical integration

needed

to alleviate problems ofspecificityduring

transition isdifferent from

what

it was inthepast.

The

reasonto verticallyintegrateundercentralplanningwastheimperfectability

of the central planner toenforce timelydelivery of

some

inputs, leading firms to

protect themselves against such disruptions. Sachs [1993] gives the

example

of

an

industrial firm installing pigpens to raise pigs

and

insure the delivery offood

to theirworkers inthe face offoodshortages.

The

typeofverticalintegration

needed

in transition, as well aslateron,islikelyto

Disorganization 23

bedifferent. It is

most

needed

forthose inputswithahigh degreeofspecificity,

and

thus alarge potentialforseriousbargainingproblems. Pigpensarenottheanswer

anymore: markets in agricultural goods

have

eliminated the problem. But the

problem

may

now

be aspecific engine part,

which

was easy to get

under

central

planning butis

now

available fromonly

one

firm, raising issues ofspecificity

and

bargaining.

Thus, the existingstructure ofvertical integrationcan onlyplay a limited role in

transition.

And

creating

new

vertically integratedstructuresofproductionboth

hassharplimits

and

takestime.

The

argumentsherearestandard. First,chains of

productioninterlock.

When

suppliers

have

more

than

one

buyer,

which

supplier

integrateswiththebuyer?Second,ifverticalintegrationistobe achievedthrough

the purchase ofexisting firms, thebargainingproblems

we

saw

become

relevant,

this time notin the determinationofinputprices, but

now

inthe determination

ofthe purchaseprice offirms. If

n

—

1 ofthe firms in ourfirst

example

vertically

integrate, theremaininghold-outfirmhas

an

incentivetoholdoutfora large part

ofthe rents,

and

thus a high purchase price. Thus,

no

firm will

want

to be the

first tobe vertically integrated. Thissuggests

much

ofvertical integration needs

toproceed throughthe

development

of

new

activitieswithin the firm. This, like

the

development

of

new

firms, takes time.

(5) Fifth, but

we

think

most

important, the very nature of(any) transitionissuch

thatthereislimitedscopeforlong-termrelations insolvingproblems arisingfrom

specificity.

Instable

economic

regimes, such as under central planning in the Soviet

Union

an

important role in reducing inefficiencies arising from specificity,

lb

see this,

take for

example

ourthirdmodel,

and assume

thatworkers

who

decidetotake

up

a private opportunitycannot

work

forthestatefirm in the future (one

can

again

replace "workers"by"suppliers" in

what

follows).This

may

be either for

techno-logical reasons (they have to

move

to another city) or because it is the optimal

punishment

from the point ofview of the state firm.

Then,

if private

opportu-nities are temporary, workers will think twice about leaving.

More

formally, they

will use a

much

higher threshold, a higher value ofc*, for deciding

whether

to

take

up

aprivate opportunity.

The

result willbeahigherprobability thatworkers

stay,

and

higherexpected productionoverall.

In that context, the value ofc* will

depend

crucially

on

the probability that the

firm is still there

and

relies

on

the

same

workers in the future. Ifthisprobability

is low, then workerswill

behave

opportunistically, in a

way

close to that

charac-terized intheprevioussection. But bythe nature of transition, this probabilityis

indeed likelytobe lowin transition: Transition implies that

many

state firmswill

disappear,

and

that,

even

ifthey survive, they will have to

change

operations as

well as

many

oftheir suppliers.

In a neo-classical

benchmark,

if it took time for

new

more

productive firms to

arise following transition, state firms

would

continue producing until theywere

replacedby new,

more

efficient firms.

The

argument

we

havejustdeveloped

sug-gests that the

mere

anticipation that

more

efficientprivate firms will eventually

appearcan lead to areduction inproductivity in the state firms. It

may

also

ex-plain

why

the pre-official transition period in the Soviet

Union

was

character-ized by increasing disorganization

and

decreasing output, as the shortening of

Disorganization

25

some

evidence that the degree to

which

state firms felt they

had

to satisfy the

plan decreased in the Soviet

Union

with the increased likelihood of reform in

the mid-1980s.

n

This line of

argument

may

explain also

some

ofthe difference

between

the

ex-periences of

China

and

Eastern

Europe

during transition. China'slower level of

industrial development,

and

itsexplicit policyof decentralizing industryto

many

smallfactoriesimplies thatproblemsofspecificitywerelessinthefirstplace (Qian

et al. [1996]).

And

the ability of the Chinese

government

tomaintain political

control

meant

thatcentralized allocation

was

not fullyreplacedbydecentralized

bargaining.

And

China's

commitment

to maintain state firms, using subsidies if

necessary,

may

have

lengthened horizons, allowing for a larger roleoflong-term

relations

between

firms

and

theirsuppliers,

and

thus avoiding thecollapse ofstate

firms in the face of

new

private opportunities. This last point naturally takes us

to a discussion of the policy implicationsofour theory.

(6) Policy implications.

Suppose

that in a neo-classical

benchmark

(i.eabsent theeffects

we

have

focused

on

so far), absent subsidies,

many

state firms

can

be expected todisappear over

time.

The

argument

we

have developed implies that, once the implications of

specificity are taken into account, this

may

well lead to the immediate collapse

11.This argument canalso explain theopposition of industry tolaws requiring firms toprovide

workersadvance notificationof plantshut-downsand whysomefirmsdecide simplytopay their

workersseverance pay rather than providingnotice. Ellickson [1989] notes thatatthetime ofthe collapseof thenineteenth centurywhalingindustry, thenormsof cooperation thathadevolved

ofthosefirms.Shorterhorizons

may

leadsuppliers to

behave

more

opportunisti-cally, leading tothe collapse ofstate production.

By

the

same

argument, a

com-mitment

by the

government

tosubsidize state firmsfor

some

time

may

avoidtheir

immediatecollapse.12

Should

a

government

commit

to subsidizing state firms,

and

thus avoid a

col-lapseofstate productionatthe beginning oftransition?This will

depend on

the

expected path of private opportunities,

on

the time path of c in our example. If

efficientprivate firms (firmswithvalues ofcequal toorgreaterthanproductivity

in thestate firms) areexpected todevelop quickly, thena

commitment

to

main-taining state firms for along time islikely tobe overlycostly. It

may

bebetter to

accept the fall in initial output, in

exchange

forhigher production

soon

after. If

however, the

development

of

an

efficientprivatesector isexpected totake

more

time (ifcisexpected tobe high

enough

tocreateproblemsof bargaining,but low

enough

that state production dominates private production for a long time), it

may

then bejustified to

commit

to subsidize state firmsfor

some

time,

and

avoid

their immediatecollapse in the face ofmediocreprivate opportunities.

There

are clearlyother implications ofsubsidies

we

have not taken into account

12.Thinkofsubsidies by thegovernment asaffecting the probability that the state firm will be

alive atany point in the future.

A

high enough probability will lead suppliers to increase the threshold atwhichthey takeuptheir privateopportunities. Technically, thecommitmentof the

governmentmustbesuch thatthere isa positive probability, howeversmall, thatthestate firm

survivesforever. Theargumentisstandard: if it wereknown thata firmwas going to disappear withcertainty atsomefuturepointin time, thingswouldunravel,and thefirmwouldcollapse at

Disorganization

27

here: subsidies

may

reduce the incentivesofstatefirms to adjust. Subsidies

have

to be financed,

and

through the tax channel, slow the

growth

of

new

private

firms.

And

in the end, political

economy

considerations

may

wellstilllead tothe

conclusion that the only

way

to stop subsidies is to stop

them

at once. But our

theory provides at least

one

respectable

argument

in favorof gradualism.

When

the

emergence

of efficient private firms is likely to be slow, a

commitment

by

the

government

to subsidize state firms for

some

time

may

avoid their

immedi-ate collapse,

and

avoid a large inefficient decline in output at the beginning of

transition.

5

Empirical

Evidence

As we

went

along,

we

gave anecdotalevidenceofdisorganization.Inthissection,

we

attempt togive

more

formal evidence aboutthe relativeimportance of

disor-ganization, both over time

and

across countries in transition.

To do

so,

we

rely

on

measuresofreported shortages ofrawmaterials

and

intermediate productsby

firms.

Two

remarks are in order here

What

firms

mean when

they report "shortages" is admittedly unclear.

And

it is

alsonotclear

whether

the

breakdowns

insupply

which

occurin ourthree

exam-plesshould be calledshortages.

There

are, however,

no

measuresof

"breakdowns

in supply", while there are measures of reported shortages.

And

it is plausible

that firms facing

breakdowns

in supply as inour examples

would

report

them

as

shortages. Thus,

we

arereasonably confident thatreported shortagesare adecent

We

focus

on

shortages ratherthan

some

of theother implications ofourexamples

because most alternative theoriesofthe output decline

do

not naturallypredict

shortages.13

There

is

one

exception: shortages

may

reflect the effects ofpartial

price liberalization, along the lines of

Murphy

et al. [1992]. In most republics

however, prices have

now

been

mostly liberalized,so that partialprice

liberaliza-tion probablyplays a limited role inexplaining shortages at thispoint.14

We

read the available evidence as suggesting that disorganization has played a

limited role in the major Central

European

countries,

some

role in Russia

and

the Baltics,

and

a majorrole in the otherrepublics.

(1) Since 1992/93, the

OECD

hascarriedout asurveyof firms in manufacturing

ina

number

ofCentral

European

and

Balticcountries,asking

them

aboutfactors

limiting production in manufacturing.

The

strength of these data is that they

come

fromroughly identical surveys. Unfortunately, they

do

not cover the early

period oftransition,

where

disorganization islikely to have

been

relatively

more

important.

The

evidence from these surveys is

summarized

in table 1:

Table 1. Percentage

of

firmsexperiencing shortages

of

materials.

13.Phenomena such aswagecompression in statefirmscompared toprivate firms (forexample

Rutkowski [1996] forPoland, Kollo [1996] forHungary), thecollapseof inter-republican trade,

cannibalization ofmachines, "spontaneous privatization" and the carvingup offirms, are also

consistentwithourstory. The problemis thatthey are alsoconsistent withother explanations, includingefficientreorganization.Thevolumeoftrade maywellhave been toohigh.

Cannibal-izationofmachinesor offirmsmaybeefficient, andso on.

14.Forasurvey ofprogressonprice liberalization asof 1995, seeEuropeanBankfor

Disorganization

29

min

percent

max

percent 1996-1 value

Czech

Republic 93-1 96-1 3 9 5

Hungary

92-1 96-1 6 11 5 Poland 93-3 96-1 3 6 5 Bulgaria 93-1 96-1 15 33 25

Romania

92-1 96-1 8 39 15 Latvia 93-1 96-1 21 32 21 Lithuania 93-3 96-1 13 43 17

Source:

OECD

Short term indicators, 2/1996. Business surveyannex.

Inthe threemajorCentral

European

countries,shortages of materials

have

played

a very limited role since 1992-1993.

The

numbers

are comparable to those

ob-tained from similar surveys offirms in

Western

Europe.

Numbers

from Berg

and

Blanchard [1994]

show

that,

even

in 1990inPoland (the firstyear oftransition)

,

disruptionswere not

an

importantpartof thestory.

The

proportion offirms citing

"supply

and

employment

shortages" as a limiting factor in production was

62%

in

October

1989,

37%

inJanuary

1990

(the

month

of the "big bang"), but

down

to

10%

byApril 1990.

The

evidence from those Central

European

countries

which

are doing less well

suggests a larger role for disorganization. In Bulgaria

and Romania, two

of the countrieswith the largest drop inoutput, supply shortages played

an

important

role

more

than

two

years after the beginning of transition.

The

same

is true of

(2) Table 2givescorresponding

numbers

forRussia, since 1991 (fromasample of

about

500

industrial firms,witha response rate of

30-40%

over time).

Table 2. Proportion

of

firmsin industry

mentioning

the following constraints as

limits toproduction. Russia,

1991-1995

Insufficient Shortages of Shortages of Shortages of

Demand

Materials Financial Resources Labor

1991 4 82 18 28

1992 51

46

45 8

1993 41 23 62 6

1994 48 19

64

5 ;

1995 47 22

64

8

Source: Russian

Economic

Barometer,

1996

The

most

striking

number

is the proportion offirms experiencing shortages of

materials in 1991. This

however

is the last year before full price liberalization

(which took place inJanuary 1992). 15

The

numbers however remained

high in

1992,

and

arestill high bystandards ofmarket economies.

Another

piece of evidence,

which

speaks

more

directly todisorganization, isthe

proportion of

work

time infirms lost to full stoppages of production, by reason.

The

time seriesare givenin table 3.

15.Itwould beinterestingtofindouthowtheproportion offirmsexperiencing shortages evolved pre-1991.Thesurvey usedabovestarts in 1991.

We

havesearched forearlierones,sofarwithout

Disorganization 31

Table3. Percentage

of

working hourslost

due

tofullstoppage

of

production,

and

reasonsforit. Russia, 1992-1995.

Percentage

By

reason

ofhourslost

Lack

of

demand

Lack

of materials

Other

1992 10 59

40

1

1993 10

50

39 11

1994 18 53 35 13

1995 7 37 53 10

Source:

Goskomstat

Rossii, variousissues.

1992

is

August

only.

1993

is the

aver-ageover 4quarters.

1994

is theaverage of

May

and

October.

1995

istheaverage

of

thefirst

two

quarters.

Lack

of materials appears to be the

main

factor in explaining

work

stoppages

during the period 1992-1995.

(3) Shortages of materials appear to play still a

more

important role in the

re-publics of the former Soviet Union. Table 4gives

some numbers

for the Kyrgyz

republic.

The

numbers

are from a survey offirms by the

ILO

and

are for 1993,

thus a year after price liberalization in the Kyrgyz republic,

and

one year after

the collapse oftrade

between

the Kyrgyz republic

and

the other republics ofthe

former Soviet Union. (1993 is

however

a period of very high inflation

—

23%

a

month

on

average

—

.

High

inflation is typically associated with disruptions in

Table 4-

Main

businessproblems offirmsin industry. Kyrgyzrepublic,

1993

Shortages of materials

Low demand

Other

factors

Food

processing 57 13

30

Textiles 36 17 47

Non

metallic minerals 26 32 42

Fabricatedmetals 31 22 47

Wood

and

paper 63 16 21

Other

43 21

36

Source: Windelletal. [1995] Proportion

of

firms, inpercent.

These

numbers

are insharp contrast to those for, say Polandor Hungary.

Short-ages of materials

seem

tobeing playinga central role.

We

are

now

collecting

more

evidence from the other republics.

So

far, the evidence appears consistent with

that inTable 4.

The

shortages inTable 4

seem

to reflect mainly internal disorganization, rather

than the collapse of trade

between

republics.

The

same

survey gives for each

industry the proportion of materials

coming

from domestic

and

non-domestic

sources.

The

correlation across sectors

between

the proportion offirms citing

shortages

and

theshareofmaterials

coming

fromnon-domesticsourcesisroughly

equal to zero.

Can

one

explain the differences

between

Central Europe

and

the former Soviet

Union?

Our

discussionsuggests a

number

ofpotentialexplanations. Specificities

Disorganization 33

say, Polandor the

Czech

Republic, leading toa larger collapse ofboth intra-

and

inter-republican trade. Distance from the West,

and

thus the

volume

oftrade,

and

thescope for foreign firms to alleviate problems ofspecificity,

may

also have

played

an

importantrole.

These

are

however

speculations at thispoint.

6

Conclusions

Transition was conceptualized in a recent

World

Bank

World

Development

Re-port[1996] as a

movement

"From

Plan to Market."Ifour

argument

is right,then a

more

accurate, iflesssuccinct, title

would

have

been

"From

Plan

and

Plan

In-stitutions to

Market and Market

Institutions."Transition

may

cause the existing

organization ofproduction tocollapse, leading to a largedecline inoutput until

new

institutions

can

becreated.

The

evidence suggests that disorganization has played a limited role in Central

Europe,a

more

importantone intheBaltics

and

Russia,

and an even more

impor-tant

one

in theother Republics. Putanotherway,disorganization appearsto

have

playeda

more

importantrole incountries

where

the output declinehas

been

the

largest.

While

political

economy

and

other considerations

may

well lead to the

conclu-sion that the only

way

to stop subsidies is to stop

them

at once, our analysis

provides

some

support for gradualist policies in dealing with state firms. Ifthe

emergence

of

new

efficient private firms islikely to be slow, there

may

be a case

References

Atkeson, A.

and

Kehoe, E, 1992, Social insurance

and

transition,

mimeo.

Banerjee, A.

and

Newman,

A., 1993, Occupational choice

and

the process of

development, Journal

of

Political

Economy

101(2).

Berg, A.

and

Blanchard,

O,

1994, Stabilization

and

transition in Poland;

1990-1991, pp. 51-92, in

The

transitioninEasternEurope,

Volume

1,

O

Blanchard,

K. Froot

and

_J. Sachseds,

NBER

and

University of

Chicago

Press, Chicago.

Blanchard, O., 1996,

The

economics

oftransition in Eastern Europe,

Oxford

UniversityPress, Oxford,

Clarendon

lectures, forthcoming.

Bryant, _J., 1983,

A

simple rational expectations Keynes-type model, Quarterly

Journal

of

Economics

3, 525-528.

Caballero, R.

and

Hammour,

M., 1996,

The

macroeconomics

of specificity,

mimeo,

June

MIT

Ellickson, R., 1989,

A

hypothesis ofwealth-maximizing norms: Evidence from

the whaling industry, Journal

of

Law,

Economics

and

Organization 5(1),

mimeo

LSE.

European

Bank

for Reconstruction

and Development,

1996, Transition report,

1995.

Greif, A.

and

Kandel, E., 1994, Contract enforcement institutions: historical

perspective

and

currentstatus in Russia.