close

Вход

Забыли?

вход по аккаунту

?

Full text in pdf file - Российская экономическая школа

код для вставкиСкачать
ÐÎÑÑÈÉÑÊÀß ÝÊÎÍÎÌÈ×ÅÑÊÀß ØÊÎËÀ
NEW ECONOMIC SCHOOL
Igor Makarov
Two Essays on Barter and Corporate Governance in Russia
Woking paper #BSP/00/XX
This paper is based on a Master Thesis prepared at NES in 2000 within a research program
“Transforming Government in Economies in Transition” sponsored by Ford Foundation, project
“Corporate Governance in Russian Firms ”. I am very grateful to the project advisors Sergei
Guriev and David Brown.
Москва
2000
Makarov I. I. Two Essays on Barter and Corporate Governance in Russia./ Working
paper #BSP/00/xx/ -Moscow, New Economic School, 2000.-38p. (Engl.)
The aim of this paper is to discuss the link between barter and corporate governance. The paper
analyses on both theoretical and empirical basis two questions. The first question is the relationship
between indebtedness and barter in the environment with a weak creditor rights protection and imperfect
credit markets. It is shown that indebted firms are more likely to use barter for avoiding payments to
creditors. Negotiations between a firm and its creditors are explicitly modeled. It is shown that the debt
overhang and lack of commitment of firm's manager not to divert cash may result in a partial or complete
failure to reduce barter. The second question we address is the analysis of barter as an entrenchment
strategy in the behavior of managers. Given that barter transactions often rely on informal agreements and
imply tight inter-temporal relationship between agents, investments in barter ought to be considered as
highly human-specific ones. Therefore one might expect that managers may abuse their power and choose
to invest in barter in order to make themselves more valuable for shareholders or owners and costly to
replace. A simple model formalizing these ideas is built and its implications are tested.
Макаров И.И. Бартер и корпоративное управление в России./ Препринт #BSP/00/xx
-М.: Российская экономическая школа, 2000. - 38 с.( Англ.)
Целью данной работы является исследование взаимосвязи бартера и корпоративного
управления в России. В работе рассматриваются два вопроса. Первым рассматривается вопрос о
зависимости доли бартера в операциях фирмы от ее кредиторской задолженности в условиях
слабой защищенности прав кредиторов и несовершенных кредитных рынках. На примере модели
показывается, что фирмы, имеющие большую кредиторскую задолженность, имеют больше
стимулов использовать бартер с целью снижения выплат по долгам кредиторам.
Второй вопрос, который нас интересует, касается возможности использования бартера
менеджером компании в целях увеличения своей значимости для компании. Это наблюдение
основывается на том, что бартерные сделки часто основываются на неформальных
взаимоотношениях участников и подразумевают долговременные отношения. В результате, выбор
бартера менеджером можно рассматривать как инвестиции в человеческий капитал, в результате
которых менеджер становится исключительным человеком для компании, увольнение которого,
поэтому маловероятно. В работе приводится модель, формализующая вышеописанные идеи и
тестируются ее выводы.
ISBN
© Макакаров И.И., 2000 г.
© Российская экономическая школа, 2000 г.
Contents
CONTENTS ....................................................................................................................................3
1.
INTRODUCTION ..................................................................................................................4
2.
EXPLANATIONS FOR THE GROWTH OF BARTER ...................................................6
3.
OWNERS VS CREDITORS: THE DEBT OVERHANG MODEL ...............................9
3.1 THE SETTING .......................................................................................................................10
3.2 ASSUMPTIONS .....................................................................................................................12
3.3 THE FIRST BEST ...................................................................................................................13
3.4 NO RENEGOTIATION ............................................................................................................14
3.5 THE MODEL WITH RENEGOTIATION ......................................................................................17
3.6 BUBBLE ...............................................................................................................................21
3.7 EMPIRICAL RESULTS ............................................................................................................22
Data description.....................................................................................................................22
Empirical results....................................................................................................................24
4.
MANAGERS VS SHAREHOLDERS: THE ENTRENCHMENT MODEL..................26
4.1 THE SETTING .......................................................................................................................27
4.2 ASSUMPTIONS .....................................................................................................................28
4.3 SOLVING THE MODEL ..........................................................................................................29
4.4 EMPIRICAL RESULTS ............................................................................................................31
Data description.....................................................................................................................31
Empirical results....................................................................................................................34
5.
CONCLUSIONS...................................................................................................................35
REFERENCES .............................................................................................................................37
3
Introduction
Corporate governance deals with the rules used to align managers’ interests with
suppliers of finance. It has been a dominant policy issue in developed market
economies for more than a decade. There is a large body of both empirical and
theoretical research conducted abroad that documents peculiar problems of
corporate governance. Among them are how to protect investors against
expropriation on the part of managers, how to provide managers with good
incentives, what is the impact of corporate governance on economic growth and so
on. A good source of recent findings for OECD economies can be found in Shleifer
and Vishny, (1997); LaPorta et al. (1998), (1999). Lately, the questions of
corporate governance have become overridingly important and most hotly
discussed issues in transition economies. At the same time the scope of problems
has widen to incorporate transition specific problems such as soft budget
constraints, weak enforcement mechanisms, existence of former state sector that
should be restructured. The focus of the problems has also changed. It is stressed
that the protection of external investors instead of dispersed ones and development
of well-functioning corporate law system should become the urgent objectives of
institutional reforms for governments in transition (Berglof and von Thadden,
2000). To achieve these goals, one needs a deep understanding and a thorough
investigation of the state of affairs in place. Situation is aggravated by the fact that
most transition countries vary considerably in history and current institutional setup
that makes cross-country analysis of reforms difficult and requires a careful
consideration of mimicking policies. What might be good in one country may well
have an adverse effect in another one when some implicit factors are not taken into
account.
As far as Russia is concerned, one of the idiosyncratic features of Russian economy
4
that can not be passed unnoticed is barter. Russia is the only country among those
ones in transition that has witnessed a dramatic outburst of barter. The barter share
in transactions of Russian firms increased from 5% in 1992 to about 50% in 1998
(Aukutsionek, 1998). It is a very disappointing result from the viewpoint of welfare
economics since the use of barter relies on exchange chains and therefore requires a
“double coincidence of wants” from the two transacting agents. This, in turn, leads
to a time consuming search as well as high inventory costs in the economy
(Kiyotaki and Wright, 1989). There are also additional “indirect” costs of using
barter such as wage arrears, weakening product market competition and slowing
down enterprise restructuring (Gaddy and Ickes, 1998). Initially considered as a
natural response to high inflation, which leads to high opportunity costs of money,
barter, however, has not disappeared as the stabilization process was proceeding.
That leads to a conclusion that there must be something else behind the
phenomenon of barter. Recent literature has proposed several explanations for the
unusual persistence of barter. Below we will provide a survey of basic lines of the
arguments. While most of them undoubtedly cover much of the story, there are still
open questions to address.
The aim of this paper is to discuss the link between barter and corporate
governance. In the paper we try to study on both theoretical and empirical basis two
questions. The first question is the relationship between indebtedness and barter in
the environment with a weak creditor rights protection and imperfect credit
markets. We show that indebted firms are more likely to use barter for avoiding
payments to creditors. We explicitly model negotiations between the firm and its
creditors and show that the debt overhang and lack of commitment of firm's
manager not to divert cash may result in a partial or complete failure to reduce
barter. The second question we address is the analysis of barter as an entrenchment
strategy in the behavior of managers. Given that barter transactions often rely on
5
informal agreements and imply tight inter-temporal relationship between agents,
investments in barter ought to be considered as highly human-specific ones.
Therefore one might expect that managers may abuse their power and choose to
invest in barter in order to make themselves more valuable for shareholders or
owners and costly to replace. We build a simple model formalizing these ideas and
test the implications.
Studying the relation between corporate governance and barter in Russia seems to
be highly relevant. Although Russia recently has adopted corporate legislation,
creditors and shareholders rights are protected only in the book and the law remains
largely non-enforced as reflected by widespread reports of cheating of minority
shareholders, corporate government scandals, and violation of creditors’ rights. The
bankruptcy procedure turns out to be prohibitively costly.
Furthermore, the
absence of a sound market-oriented financial system, which could rank projects by
risks and returns and sanction managers who fail to maximize shareholder value,
allows development of quasi-financial institutions such as barter.
The paper is organized as follows. In section 2 we survey basic explanations for
barter. In section 3, we build a model of debt overhang with renegotiation and
transformation in the spirit of Hart (1995) and elements of Myers and Rajan (1998).
The implications, namely the positive relationship between outside debt and barter,
are tested in subsection 3.7. In section 4 we build a simple model of barter as a
managerial entrenchment strategy. In subsection 4.4 we test the implications of the
above model, the main of them is a negative relationship between barter and
manager’s share of firm’s assets. Section 5 concludes.
1. Explanations for the Growth of Barter
It is common knowledge that existence of the institute of fiat money crucially
6
hinges on agents’ beliefs, since money has no intrinsic value and possesses it only
in agents’ minds. Once being accepted, money gives rise to development of other
economic institutions (e.g. tax system, banking system and so on) whose role is to
help to maintain and promote economic growth or more generally human beings’
welfare. These institutions, however, relate to the institute of money not in a
unilateral direction but bilateral one. Malfunctioning of any of them may severely
undermine the value of money and desire to use it. It seems that in case of Russia it
was the inadequate work of such institutions, along with inappropriate
macroeconomic policies, that caused demonetizing and emergence of barter.
In 1992-1995, as a result of price liberalization and transition to market economy,
Russia has witnessed a period of high inflation. The usual costs of inflation in case
of Russia were exaggerated by the lack of a sound banking system. It took, for
example, about a month to pass payments from one firm to another. During that
time lap the value of money, away from indexation, often reduced in 20-30%. This
led to the situation when cost of using barter became comparable with that of using
money and this, in turn, gave rise to barter (Hayashi and Matsui, 1996). However,
this conventional explanation does not make the whole story - as inflation calmed
down barter did not fade at all.
In 1995-1998 to combat the inflation, a new stabilization program was adopted
which was primarily based on sharp tightening of monetary policy, although not
accompanied by a respective cut in government spending, and fixed exchange rate.
At the same time the source of deficit finance changed sharply to GKO market
which provided very high returns on treasury bills. As a consequence, banks shifted
their portfolios to financing the government, cutting off the real sector of the
economy from credits. In this situation firms, being liquidity constrained, were
forced to use barter as means of short term-credit (Commander and Mumssen,
1998).
7
At the same time, Russia had some institutional peculiarities that fostered nonmonetary transactions. First, money in banks was not freely convertible into cash.
There was a system of non-cash transactions between firms within the banking
system when firms were permitted to convert non-cash into cash only in a limited
number of cases. Thus, firms had limited access to cash outside the banking system.
The need for outside transactions was mainly due to the bizarre tax system, which
was the second factor playing important role in burgeoning of barter. According to
Russian tax law, all tax payments must go through the banking system. The use of
barter allowed firms to circumvent banks, shadow its financial streams and thus
evade taxes. The role of this strategy was ambiguous. While some authors argue
that it could explain a large part of the barter story (Hendley, Ickes and Ryterman,
1998; Yakovlev, 1999b) others do not think so. Commander and Mumssen (1998)
show that most firms believed that barter rather increased their tax bills, partly
because the barter prices almost always exceeded cash prices.
When an enterprise is in arrears in its tax payments, tax officials can block its bank
accounts, forcing all of the firm’s current revenues into a special account that was
applied to the payments of its tax debt. This procedure often took place even if a
firm could not pay its tax liabilities because of arrears by the state. A firm in tax
arrears must pay a penalty about 0.3% per day, so very soon it becomes hopelessly
insolvent. As a result, in 1995-1998 most of the Russian enterprises were in tax
arrears and had mutual debt liabilities. In this situation barter represented a survival
strategy. It allowed a firm to operate outside the banking system and maintain its
output1.
While in most countries with developed economic institutions enterprises with debt
liabilities and tax arrears have few chances to survive, in 1995-1998 in Russia
1
Thus barter served, in a sense, as a defense mechanism and its existence may, at least partially,
explain the fact that Russia has not suffered much from the banking crisis in August 1998.
8
situation was the opposite. First of all, bankruptcy procedures turned down to be
hardly implemented because there was a fear that liquidation of a small number of
key enterprises (basic debtors) could initiate a long chain of bankruptcies and firms
generally did not believe that the claims of secured creditors have priority over
government claims against enterprises. Second, due to political games and
interference between federal and regional governments enterprises often had an
opportunity to pay their tax liabilities via “soft goods”. As a consequence, a lot of
former Soviet Union’s firms chose to invest in barter that provided them with a
greater flexibility in bargaining process with the state (Gaddy and Ickes, 1998).
Thus, while for firms in a good situation but without access to bank credit barter
was a substitute for short-term credit, for indebted firms barter was rather a way to
avoid costly restructuring and source of external finance (Brana and Maurel, 1999).
Weak corporate governance nurtures such opportunistic behavior also. Managers
often use barter to hide transactions from outside owners in order to increase their
ability to control the enterprises (Ickes and Ryterman, 1997).
2. Owners Vs Creditors: The Debt Overhang Model
In this section, we suggest a simple model that describes behavior of a liquidity
constrained firm with an outside debt. Essentially, the model is the one of a debt
overhang. The firm faces the following choice: if it pays the debt off, it will be
stripped of the working capital and will not be able to purchase inputs for the next
round of production. Therefore the firm would rather prefer to hide its revenue.
One way to do this is to sell for barter that cannot be expropriated by the creditor
(or has no value to the outside creditor). Even if barter is costly, it allows to
postpone the debt payments, finance another round of production which may then
9
allow to pay off the debt.
This story is a conventional explanation of the link between the liquidity constraint
and barter (see, for example, Commander and Mummsen (1998)). The story,
however, is not fully consistent. Indeed, if the firm is efficient and each additional
round of production add value, why would not the creditor voluntarily restructure
the debt? Since forgiving/refinancing the debt allows to increase utilization of
efficient capacities of the firm, voluntary renegotiation would result in the
postponement of the debt service. This argument is common in literature
concerning financial contracting in the developed economies (Hart (1995), Ch. 5)
and on the debt relief for developing economies (Krugman (1988)) so that it is not
at all clear why it should not apply to a transition economy.
We provide two alternative answers to this puzzle. First, it turns out that even in the
presence of renegotiation, the risk of cash diversion by the firm manager (the
transformation risk) may cause barter to emerge. Second, if the creditor has an
access to investment opportunities that the firm does not have and these
opportunities yield very high returns (like it were in the case of Russia's
government bond bubble), the creditor will not be interested in the debt
restructuring.
One of the cornerstones of our model is the lack of effective bankruptcy
procedures. Unlike the conventional models of debt (Hart (1995)), we assume that
the creditor cannot get control over the firm's assets if the firm breaches the
contract.
2.1 The setting
There are two agents: a firm F and an outside creditor C. F owes C debt D0 > 0. The
10
F sells m0
for cash
t=1 F buys x units
of input for cash
t=0
F owes C
debt D0
F sells output for
cash or barter and
pays D1
F and C
negotiate
P and D1
F sells b for
barter, gets βb
units of input
F produces
λ (x+βb)
units of output
Game
ends
Figure 3.1. Timing.
firm has no cash and has a unit of output. The firm may sell the output either for
cash or for barter and uses the revenue to purchase inputs. The relative prices in the
cash market are better for the firm than those in the barter market: the barter
transactions involve high legal, transportation and storage costs. On the other hand,
the cash revenues can be captured by the creditor while the barter ones cannot. The
cash revenues are accrued to F’s current account which the creditor can easily
seize. The in-kind payments have no value to the creditor and can only be used as
inputs in F’s production.
There are two periods. The timing is as in the Fig.3.1. First, F chooses share of
output to be sold for cash m0 and the share of output to be sold for barter b
(m0+b≤1). The cash prices of output and input are normalized to 1: selling m0 for
cash, F gets m0 rubles that can buy m0 units of input. The relative barter prices for
inputs are β∈[0,1] where (1-β) represents the transaction costs of barter. Thus,
selling b for barter, F gets βb units of input. At time t=1, the debt is due. F and C
observe F’s cash revenues and can renegotiate the contract. They bargain over a
new contract (P,D1) where P are the payments at time t=1 and D1 is the new debt
11
due at t=2.2 If the renegotiation succeeds, F’s cash balance becomes m1 =m0 – P,
and F promises to pay D1 at t=2.
If the renegotiation fails, C takes P=min{m0,D0} and invests it elsewhere. In this
case, F only has m1 =m0 – min{m0,D0}=[m0 – D0]+. The new debt is
D1 = D0 – min{m0,D0}=[D0 – m0]+
After the renegotiation, F buys inputs for cash. If the firm spends x∈[0,m1] rubles
on buying inputs, the total amount of inputs the firm can use for production is
q = βb + x. The firm has a linear technology that converts q units of input into λq
units of output. The maximum capacity is one unit of input: q≤1. The capacity
constraint never binds since q ≤ b + x ≤ b + m1 ≤ b + m0 ≤ 1.
Once the output λq is produced, F decides again whether to sell it for cash or for
barter. The cash revenues m2 ∈ [0, λq] can again be confiscated by the creditor if
the debt has not been repaid yet. The remaining cash is used for consumption by
F’s owners. Then the game ends.
The gross interest rate in the economy is normalized to 1. Therefore the creditor’s
and the firm's payoffs are
UC=P+min{m2,D1}
(1)
and
F
U =m1–x+[m2–D1]+,
(2)
respectively.
2.2 Assumptions
1. For simplicity's sake we will make a few assumptions on the firm's productivity
λ and transaction costs of barter 1–β:
1 < 1/β < λ < 2.
The first inequality implies that barter is less efficient than money: β < 1. The
2
The existing contract gives C a right to claim D0 but F cannot physically pay more than m0.
12
second inequality states that the firm's productivity is high enough so that even
at the relative prices β, the firm adds value: λβ > 1. Together, these two
inequalities imply that the firm adds value at the cash prices as well: λ > 1. The
last condition is a technical one and makes the problem less trivial. If the firm
were too productive λ > 2, the renegotiation would always postpone the debt
service. The gains of another round of production were so large that they would
always overcome the transformation risk (i.e. F's diversion of cash m1 for
current consumption).
2. F has all bargaining power. This is again a simplifying assumption. We give all
the bargaining power to the firm in order to show that even if F is the residual
claimant, cash may still be expropriated which would provide incentives for
barter.
3. Parties have symmetric information. Cash has the same value to both parties.
Barter can only be used as an input in production using the technology owned
by F. The technology is inalienable. Even if F breaches the debt contract, C
cannot take control of the productive assets.3
2.3 The first best
Apparently, the social optimum is to sell for cash, buy one unit of input and
produce at maximum capacity. In other words, b=1, m0=1, q=1, m2=λq. The social
welfare equals λ.
One way to achieve it is to set P=0 and D1=0 which is equivalent to debt
forgiveness. Certainly, this contract does not meet the creditor's individual
rationality constraint.
3
The conventional explanation of this assumption is inalienability of human capital. Although it may be applicable
to Russia, we have in mind much bigger problem: the absence of effective bankruptcy procedures. In Russia,
creditors have hard time claiming the assets of bankrupt firms.
13
There are three potential sources of inefficiency in the model. First, selling for
barter rather than for cash in the beginning. Barter sales involve transaction costs
(1–β)b. This inefficiency could be dealt with effective bankruptcy procedure. In
this case if the debt is not paid on time, C would take control over production
technology. Therefore F gains nothing by using barter.
The second source of inefficiency is the failure of the debt renegotiation: if C takes
all the cash F has at t=1, F produces below social optimum q<1, and therefore the
deadweight loss (λ-1)(1-q) arises. Similarly, the third problem is the transformation
risk: even if the debt payments are postponed and F keeps some cash, F may prefer
to spend it for consumption right away instead of purchasing inputs. As in Myers
and Rajan (1998) we assume, that the more liquid the asset is the higher the
transformation risk is: cash can be transformed, while barter can only be used for
production. The transformation risk often appears in the literature on incomplete
financial contracts where the only contractible variable is the payment from one
party to the other one, while the levels of input and output are not contractible.
2.4 No renegotiation
Let us first study what happens if the renegotiation at t=1 is not allowed. We will
solve the model via backward induction. First, we will find m2 under given x, P, D1,
m0 and b. Second, we will determine x given P, D1, m0 and b. Then we will find P
and D1, given m0 and b. Finally, we will describe the choice of m0 and b. The
ultimate goal of our analysis is to check how the choice between money and barter
depends on debt D0.
The choice between money and barter at t=2 is trivial. Since barter can only be
used for production, it makes no sense to sell for barter.4 Apparently, m2=λq.
4
Certainly, this is due to the finite horizon setting.
14
The amount of inputs bought for cash x ∈[0, m0 – P] is chosen by the firm to
maximize
UF = m0 – P – x +[λb(1-β) +λx– D1 ]+.
This function is convex with regard to x. Therefore we have the corner solution:
either x = 0 or x = m0 – P. F uses all available cash x = m0 – P to buy inputs
whenever [λβb +λ(m0 – P)– D1]+ ≥ m0 – P +[λβb – D1]+. Since m0 – P ≥ 0, this
condition is equivalent to
λβb – D1+(λ–1)(m0 – P) ≥ 0.
(3)
Let us now determine the first-period payments P and the second period debt D1.
Since there is no renegotiation, P =min{m0,D0} and D1=[D0 –m0]+.
Now we shall describe the choice between money and barter. The firm chooses b
and m0 to maximize (2) subject to b+m0≤1. There can be 2 cases:
1. The firm gets enough cash revenues to pay off the debt: m0≥D0. In this case P
=D0 and D1=0. Inequality (3) holds, so that x= m0 – P and the firm's payoff (2)
becomes UF = λβb +λ(m0 – D0). Since β<1, the firm is better-off selling for cash
as much as possible: m0=1 and b=0. Obviously, this case is only possible if
D0≤1. The F's utility is UF m=1= λ(1 – D0).
2. The firm sells very little for cash: m0,> D0. The firm pays all the cash to the
creditor: P =m0, and D1=D0 –m0,. There is no cash to buy inputs x=0 and the
firm's payoff (2) becomes UF = [λβb + m0 – D0]+. Since λβ>1, the firm is betteroff selling everything for barter: m0=0 and b=1. Thus the firm gets
UF b=1= [λβ – D0]+.
Comparing the two cases we see that the firm is better-off selling for barter
15
whenever the debt is sufficiently high (see Figure 3.2).
Proposition 1. In the model without renegotiation, the firm chooses to sell all its
output for barter UF b=1 ≥UF m=1 if and only if D0≥ D*=λ(1-β)/(λ-1). Otherwise, the
firm sells all its output for cash.
UF
UFm=1
UFb=1
0
D*
1
λβ
D
Figure 3. 2. Firm's payoff as a function of outside debt in the model without
renegotiation.
The Proposition is quite intuitive. Indeed, if there is no renegotiation, the creditor
will seize all the cash the firm gets for its sales. Stripped of the working capital, the
firm will not be able to continue production at a reasonably high level. To protect
its working capital the firm chooses to hide the income from the creditor via selling
for barter. Although inefficient (β<1), barter allows to avoid expropriation of the
working capital and ensure a possibility to buy inputs. Being able to produce in the
second period, the firm gets cash and partially or fully pays off the debt.
16
2.5 The model with renegotiation
The model above assumes that the renegotiation is not allowed and the creditor
seizes all the cash the firm has at t=1. Apparently, this is a myopic behavior: by
restructuring the debt, the creditor would encourage the firm to produce more in the
next period. This, in turn, would increase the creditor's chances to get the money
back. Debt restructuring may therefore provide the ex ante incentives for the firm to
sell cash in lieu of barter. F knows that C will not expropriate all the cash right
away, so there is no need to hide the revenue in the form of barter. In this section,
we study a model with renegotiation and check to which extent renegotiation helps
to reduce barter.
Again, we will solve the model via backward induction. Apparently, the choice of
m2 and x under given P, D1, m0 and b is the same as in the previous section. In the
second period, F sells for cash: m2=λq. The amount of inputs bought for cash is x =
m0 – P whenever (3) holds and x = 0 otherwise.
Now let us consider the debt renegotiation. The firm and the creditors bargain on P
and D1 to maximize joint surplus of F and C at the date t=1. At this point, the
choice between money and barter m0 and b has already been made, so that the
renegotiation only affects the last-period debt overhang and therefore F's incentive
to produce more or less. If P and D1 are such that (3) holds, then (3) F uses the
remaining cash if any to buy inputs and produce more. Otherwise F spends the
remaining cash right away and only uses inputs bought for barter. The parties'
payoffs calculated at t=1 are as follows:
1. Inequality (3) holds:
UC =P + min{λβb +λ(m0 – P), D1}; UF =[λβb+λ(m0 – P)– D1]+.
2. Inequality (3) is violated:
17
UC = P +min{λβb , D1}; UF=m0 - P +E[λβb – D0 + P]+ .
Since F has all the bargaining power, F chooses P and D1 to maximize UF subject
to the creditor's participation constraint UC ≥ m0 + min{λβb , D0 – m0}.
It is easy to show that the case 2 is suboptimal. The idea of renegotiation is to
postpone the debt payments to leave some cash for input purchases. If the firm uses
the cash for current consumption x=0, then there is no point in renegotiation. Thus,
the second case will never occur in equilibrium. To make sure that it does not
happen, the parties will agree on a contract P, D1 that satisfies the constraint (3) and
therefore prevents the firm from diverting the cash. Unfortunately, the constraint
(3) may be binding which may distort the choice of P , D1: the diversion never
happens but the threat of diversion may prevent the parties from achieving the first
best.
Formally, F chooses P , D1 to maximize
[λβb+λ(m0–P)–D1]+
(4)
subject to (3) and the creditor's individual rationality constraint
P+ min{λβb +λ(m0 – P), D1 }≥ m0 + min{λβb , D0 – m0}.
The solution to this problem is P=(2–λ)-1[m0 + min{λβb , D0 – m0}–λβb–(λ-1)m0]+,
D1 = m0 + min{λβb , D0 – m0}–P. Thus, whenever
18
m0 + min{λβb , D0 – m0} > λβb+(λ-1)m0, the renegotiation results in P>0, some
cash is used to pay off the debt rather than to buy inputs. This provides F with
wrong incentives: indeed, selling for cash, F would lose some of its working capital
and therefore will not be able to produce as much as it could if it would have sold
its output for barter. Notice that P>0 if and only if constraint (3) is binding – if
there were no risk of cash diversion, the parties would postpone all the debt
payments P=0.5
Substituting P, D1 into (4), we obtain the firm's payoff as a function of b and m0
UF = m0+([λβb +m0 –D0]+– (2–λ)m0)–(2–λ)-1(λ–1)[(2–λ)m0 – [λβb +m0 –D0]+]+ (5)
UF
UFm=1
UFb=1
0
λ-1
λβ
D** 1
D
Figure 3.3. Firm's payoff as a function of outside debt in the model with renegotiation.
5
This is precisely where the assumption λ>2 kicks in. If we allowed for λ<2, the condition (3) would never bind,
and the solution would always be P=0, D1 = m0 + min{λβb , D0 – m0}. Therefore, renegotiation would eliminate
barter altogether.
19
The firm chooses b and m0 to maximize (5) subject to b+m0≤1. Apparently, (5)
increases with both b and m0, so that this constraint binds b+m0=1. Substituting
b=1– m0, we obtain a convex function of m0. Therefore the solution is always a
corner one: either sell everything for cash m0=1 or sell everything for barter m0=0.
If F sells everything for barter it gets UF
b=1=[λβ–D0]+.
If F sells everything for
cash, it gets UF m=1= 1+([1 –D0]+– (2–λ))–(2–λ)-1(λ–1)[(2–λ)–[1 –D0]+]+.
Figure 3.3 shows UF
b=1
and UF
m=1
as functions of the initial debt D0. One can
easily show that UF b=1 > UF m=1 if and only if D0 > D**=(1–λβ(2–λ)) / (λ–1).
Proposition 2. In the model with renegotiation, the firm chooses to sell all its
output for barter UF
b=1
≥UF
m=1
if and only if D0≥ D**=(1–λβ(2–λ)) / (λ–1).
Otherwise, the firm sells all its output for cash. In the presence of renegotiation,
barter is less likely: D*<D**.
The Proposition implies that introduction of renegotiation makes barter less likely:
if D0 ∈(D*, D**) the barter does not occur in the presence of renegotiation but it
would occur if the renegotiation were not allowed.
Thus, renegotiation reduces barter but does not eliminate it altogether. Why is this?
The key is the transformation risk: in the debt renegotiation the firm's manager
cannot commit not to divert the cash for current consumption in preference to
purchasing inputs. The reason why the firm's manager may be interested in
diversion is the remaining debt overhang. If the debt is rescheduled, and the
second-period debt burden is too high, the firm expects getting too little of the cash
revenues m2. Therefore diversion is likely to happen. Diversion can be prevented by
reducing the second period debt overhang. But this is costly: to compensate the
creditor for the lower second-period return, F has to pay more in the first period
which, in turn, provide F with the wrong ex ante incentives: F knows that it will
20
have to pay something in the first period and will prefer to have as little cash and as
much barter as possible.
2.6 Bubble
Another important explanation of the failure of renegotiation to eliminate barter is
the high real interest rate. In 1995-98, Russian government bond (GKO) market
paid off very high real returns. Since the bubble busted in August 1998, the level of
barter has been steadily decreasing. We will introduce the GKO market in the
following way.
Suppose that the creditor has an investment opportunity that pays off a gross
interest rate δ>λ. The firm does not have access to this opportunity. At the time of
renegotiation the parties expect the following payoffs: UC=P+δ-1min{m2,D1} and
UF=m1–x+[m2–D1]+.
Solving the model by backward induction, we obtain the equilibrium which is
equivalent in real terms to the equilibrium without renegotiation. Indeed, m2=λq,
x= m0 – P if (3) holds and x=0 otherwise.
The renegotiation ends up with P =min{m0,D0} and D1=[D0 –m0]+. Indeed, C is not
interested in the second period payments unless F offers δ second-period rubles for
each first-period one. But this does not happen: F's internal rate of return is λ. Thus,
the choice between money and barter at t=0 is precisely the same as in the model
without renegotiation.
Certainly, stripping the firm of its working capital and buying the bonds are locally
efficient: the coalition of F and C makes more money via investing in GKO rather
than through buying inputs and producing. We should not however forget that δ is
not a market rate of return. Rather, it is a growth rate of a bubble. The cost of
capital in the economy is still normalized to one, and therefore redirecting the cash
21
from the real sector to the bond market is not efficient for the economy as a whole.6
2.7 Empirical results
Data description
The model in Section 3 predicts positive relationship between outside debt and
share of barter in sales. For testing the above model we use Goskomstat data and
the dataset Barter in Russian industrial firms built in the New Economic School
Research Project « Non-Monetary Transactions in Russian Economy ». This
dataset was created by matching the surveys of managers of Russian industrial
firms conducted in 1996-98 by Serguei Tsoukhlo (Institute of Economies in
Transition, Moscow) with Goskomstat database of Russian firms (Federal
Committee of Statistics of Russian Federation). Since Goskomstat data were most
complete for 1996 and 1997 we ran regressions for 1996 and 1997 data.
The barter data included six to seven hundred firms each year. The barter data are
answers of firms' managers to the following (eight) questions: « how much of your
firm's inputs (outputs) were paid in rubles, in dollars, in kind and in promissory
notes? » The Goskomstat database includes compulsory statistical reports that all
large and medium-size firms must submit to Russian Federal Statistics Committee.
There are over 16 thousand firms in the database. However, there are many missing
items in the financial accounts. After matching barter data with the Goskomstat
data we ended up with roughly three hundred observations in each year among
which only about a hundred firms appears both in 1996 and 1997.
6
Certainly, this is not a closed and consistent model with rational players: we just look at one end of the GKO cash
flows. Building a general equilibrium of this Ponzy game is however not the purpose of this paper. We may only say
that somewhere in the public or private sector, there were irrational traders who supported the high rate of return.
22
As proxy of debt we take the firm's total indebtedness in the beginning of the
corresponding year divided. This variable (sum of line items 610 and 620 in the
balance sheet) includes bank loans (610) and amounts owed to suppliers,
subsidiaries, consolidated government, IOU holders, employees and other creditors
(620).
To control for other explanations of barter we also include in our regressions firms'
size, export and market power, regional and industry dummies. As a proxy of size
we take logarithm of annual sales. As for export, we include share of export in
sales. The concentration ratios CR4 (share of four biggest firms in total sales of an
industry) are calculated for 5-digit OKONKh industries (more than three hundred
industries) using the Goskomstat database. The summary statistics and pairwise
correlations are provided in the Tables 3.1 and 3.2 below.
Table 3.1
Variable
Explanation
Mean
Std. Dev. Min
Max
Barter_96
Share of barter in
sales, 1996
Share of barter in
sales, 1997
Debt as of Jan 1,
1996 divided by
annual sales
Debt as of Jan 1,
1996 divided by
annual sales
Log sales 1996
Log sales 1997
Share of export in
sales, 1996
Share of export in
sales, 1997
CR4 in the 5-digit
industry in 1997
CR4 in the 5-digit
industry in 1997
0.37
0.24
0
.83
0.42
0.25
0
.83
0.28
0.48
0
5.07
0.61
2.02
0
31.6
17.0
17.0
0.084
1.72
1.79
0.166
11.1
9.1
0
22.3
21.6
.97
0.063
0.146
0
.97
0.37
0.26
0.044
1
0.38
0.26
0.037
1
Barter_97
Debt96
Debt97
ls96
ls097
export96
export97
Cr496
crou497
23
Table 3.2 : Pairwise Correlations
Barter_96
Debt96
Ls96
Export96
cr496
Barter_97
Debt97
ls97
Export97
cr497
Barter_96
Debt96
Ls96
Export96
cr496
1
0.25***
0.11
0.10
0.26***
1
0.10
0.27***
0.22***
1
0.1836
0.34***
1
0.32***
1
Barter_97
Debt97
ls97
export97
cr497
1
0.30***
0.09
-0.02
0.14**
1
0.06
0.13**
0.13**
1
0.20***
0.39***
1
0.25***
1
We also included 2-digit industry dummies as defined in the Table 4.3 below. As
for the regions we have only introduced dummies for Moscow (msk), Urals (ural)
and Siberia (asia) with the base category being European Russia.
Table 3.3 Composition of IET survey sample by 2-digit industries
Ind1
Ind2
Ind3
Ind4
Ind5
Ind6
Ind7
Ind8
Ind9
Ind10
Ind11
Industry
Electricity
Fuel
Ferrous metals
Non-ferrous metals
Chemical and petro-chemical
Machinery
Pulp and forestry
Construction materials
Textile
Food
Other
Number of firms, 1996
2
10
44
13
78
166
54
64
89
86
15
Number of firms, 1997
2
7
40
12
66
203
106
59
83
89
14
Empirical results
We run OLS regressions for barter for 1996 and 1997. The results of the
24
estimations are shown in the Table 3.4. It is seen that coefficient at debt comes out
positive and significant.
In addition to running the regressions for the whole sample, we estimate the effect
only for the firms with indebtedness below annual sales. These are the firms that
are not completely hopeless and therefore our model should be more applicable to
them. Indeed, the effect of debt on barter is indeed stronger and more significant for
those firms.
Table 3.4 : OLS regressions for barter
Barter
Debt
Log sales
Export
1996
whole sample
0.07** (0.03)
.016 (0.011)
-0.21*** (0.07)
cr4
Msk
0.13* (0.07)
-0.13*** (0.04)
Ural
Asia
Ind2
Ind3
Ind4
Ind5
Ind6
Ind7
0.08* (0.05)
0.06 (0.04)
0.11 (0.13)
0.22** (0.09)
0.07 (0.12)
0.22** (0.09)
0.21** (0.08)
0.36*** (0.09)
Ind8
Ind9
Ind10
Constant
0.18** (0.09)
0.16* (0.09)
0.01 (0.09)
-.01 (0.21)
1996,
Debt96<1
0.12** (0.06)
0.01 (0.01)
-0.23***
(0.08)
0.11 (0.08)
-0.13***
(0.04)
0.10* (0.05)
0.06 (0.04)
0.12 (0.13)
0.22** (0.09)
0.07 (0.13)
0.23** (0.09)
0.21** (0.08)
0.35***
(0.09)
0.17** (0.09)
0.14 (0.09)
0.01 (0.09)
-.03 (0.22)
N
R2
290
0.32
276
0.32
1997
whole sample
0.03** (0.01)
0.01 (0.01)
-0.32*** (0.08)
1997, Debt97<1
-0.05 (0.06)
-0.22*** (0.04)
-0.08 (0.06)
-0.22*** (0.04)
0.14*** (0.04)
0.08** (0.04)
-0.06 (0.15)
0.24** (0.11)
0.05 (0.15)
0.22** (0.10)
0.17* (0.10)
0.22** (0.10)
0.16*** (0.05)
0.09** (0.04)
-0.06 (0.14)
0.22** (0.10)
0.05 (0.14)
0.21** (0.10)
0.14 (0.09)
0.21** (0.10)
0.10 (0.10)
0.20* (0.10)
-0.08 (0.10)
0.17 (0.19)
0.10 (0.10)
0.15 (0.10)
-0.10 (0.10)
0.21 (0.19)
316
0.34
271
0.38
*** - significance at 1% level, ** - significance at 5% level, * - significance at 10% level
25
0.17*** (0.04)
0.01 (0.01)
-0.38*** (0.09)
3. Managers Vs Shareholders: The Entrenchment Model
In this section we build a simple model that analyzes the new set of entrenchment
strategies to which a Russian manager has an access. It was mentioned earlier that
the most striking feature of Russian economy over the last decade was barter. As
was stressed by many authors (Kiyotaki and Wright, 1989; Polterovich, 1998)
individual rationality of using barter depends to a certain degree on to what extent
other agents agree to use this strategy. Given widespread use of barter in Russia,
one may suppose that the cost of using barter here is not so high as this is in
countries with little share of barter transactions. At the same time, investment in
barter is a human-specific one. What is more, the return from barter often is very
uncertain for outsiders since it crucially depends on informal relationships and
agreements (Yakovlev, 1999a). As a result, one can view barter as a good
managerial entrenchment strategy, moreover, the use of barter could be easily
justified by hard economic situation of a partner or even by presumed benefits
from a barter deal.
As a backbone of our model we use the framework of Edlin and Stiglitz, (1995)
with some elements of Shleifer and Vishny, (1989) applied to barter. In Edlin and
Stiglitz, (1995) to entrench himself a manager invests in the project with
asymmetrically observed return. While the return is certain for the incumbent
manager, other economic agents perceive it as a random variable. Being risk-averse
then competitive managers demand higher compensation for their jobs than the
incumbent manager does. As a result, ceteris paribus the manager in office will
have clear advantage over its rivals. In Shleifer and Vishny, (1989) the key element
is the manager’s investment in the project he can run best. Thus he becomes more
valuable for the firm and has better chances to keep holding the office. In the model
we combine these two effects, which, clearly, work in the same direction in case of
26
barter. Among exogenous parameters of the model there is a share of the firm’s
assets the manager owns. We show that as the share grows the manager is less
likely to use barter.
3.1 The Setting
There are three principal agents: an incumbent manager (I) who runs the firm’s
asset, a board of directors (B) who controls the manager and a representative
competitive manager (R - rival).
Figure 4.1. Timing.
t=0
I chooses investment
levels m for cash and b
for barter
t=1
R makes an offer
B decides to accept
or decline it
t=2
Profits are
realized. Players
get their returns
Game ends
There are two periods. The timing is as in Fig. 4.1. Initially the firm has one unit of
asset and two projects to invest. The first project has return 1 on 1 unit of
investment, while the second one has only β < 17. In the first period the manager
chooses investments. Investments are supposed to be irreversible, that is, once
made they cannot be recovered by reselling them. We suppose that only the second
project requires manager-specific investments. Moreover, information about the
return on the second project is asymmetric: the return is certain for the insiders and
noisy for the outsiders. Thus, if I chooses to invest m units in the first project and b
units into the second provided that he will manage the asset he gets βb. If R
supervises the asset, he expects to get K(m+βϕ(b))+ µ(b)θ, where θ is a random
variable (only for R) with zero mean. Here K measures the ability of R to run the
assets, the term ϕ(b) reflects the fact that investments are manager-specific (it is
7
Again we assume that interest rate is normalized to one.
27
assumed that ϕ(b)/b is a decreasing function) , µ(b) is an increasing function of b
because the noise increases with b.
Remark: As was argued above, one may identify barter with the second project
while the alternative first project means investment in new product lines or in
quality of products.
At the time t=1 R makes his offer demanding a compensation for his efforts. Given
the investments, the board of directors considers this offer. If the gains outweigh
those which the company can achieve under the incumbent manager the offer is
accepted, I is fired and R becomes a new CEO of the firm. Otherwise the offer is
declined.
At the time t=2 profits from investments are realized. Players get their returns and
game ends.
3.2 Assumptions
1. We assume that remuneration of a manager consists of a fixed wage income
w and share of profits s. Only w is subject to bargaining8. This linear
compensation scheme being rather simple in nature allows to make a quite
general analysis and get rich conclusions.
2. Both incumbent and competitive managers have CARA utility functions with
absolute risk aversion parameter 2r. The distribution of the random variable
θ is N(0, σ2). These are simplifying assumptions. The analysis does not
change much if we take another utility function or distribution of θ.
However, this specification is useful, since it highlights the role of risk
aversion and noisy nature of barter.
8
Because capital markets in Russia are imperfect. That results in rigid ownership structure.
28
3. The outside opportunities give R return C. We assume that there is a
competitive managerial market. As a result, R will demand w such that
expected return equals C.
4. The rival’s ability to manage assets K is a random variable with distribution
function G (publicly observed). This assumption stems from the simple fact
that in reality people differ in their abilities. For the incumbent manager K=1
is a normalization.
5. We also assume that the greater the share s, the lesser the possibility that the
incumbent manager will be fired. To take it into account we introduce an
increasing in s function of dismissal - ρ(s).
3.3 Solving the Model
We will solve the model via backward induction. First, given investment levels m
and b the distribution of R’s ability K determines the probability of I’s dismissal.
Second, given this probability we will find the choice of m and b. We will use
subscripts I and R to denote the variables that correspond to I and R.
If the levels of investments are m and b (obviously, it must be m+b=1) R’s
expectation of return on firms assets is given by
U R = EU ( w + s ( K (m + βϕ (b)) + µ (b)θ ))
As we assume, CARA form for R’s utility function certainty equivalent CER for R
equals
CE R = w + sK (m + βϕ (b)) − rs 2σ 2 µ (b) 2
Thus, to have the income C R will require the wage wR
w R = C − sK (m + βϕ (b)) + rs 2σ 2 µ (b) 2
If I manages the asset then the firm will have the profit
Π I = (1 − s )(m + βb) − w I
If R manages the asset then the firm will have the profit
29
Π R = (1 − s ) K (m + βϕ (b)) − w R = K (m + βϕ (b)) − C − rs 2σ 2 µ (b) 2
As a result, I will be fired off if and only if
2
(1 − s )(m + βb) − w I + C + rs 2σ 2 µ (b)
Π > Π + ρ (s) ⇔ K >
+ ρ (s)
m + βϕ (b)
R
I
Since m+b=1 it is equivalent to
2
K>
(1 − s )(1 − (1 − β )b) − w I + C + rs 2σ 2 µ (b)
+ ρ (s)
1 − (1 − βϕ (b) / b)b
(1)
The greater the RHS of (1), the lesser the possibility that the incumbent manager
will de dismissed. This gives us the probability that I will not be firing
p( s, b) = 1 − G ( RHS of (1)) . As can be seen from (1) p(s,b) increases with b. We can
identify two sources of this effect. The first comes from the fact that investment in
barter is a manager-specific one. So, investing in barter I becomes more valuable to
the firm, because he is the person who suits best for the job. The second is
essentially due to asymmetric information. Since R is a risk-averse agent, he does
not like risk and requires more compensation than I. Here we suppose unlike
(Edlin and Stiglitz, 1995) that gains from investing in barter result in increase in
probability to stay within the firm but not in improving I’s bargaining position
concerning its wage. It seems to be a very natural assumption in immature labor
markets where employment rights are weakly protected and negotiations are rare.
We also make additional assumptions about the function p(s,b). Namely, we will
assume that
′′ ( s, b) < 0
p s′ ( s, b) > 0 and psb
(2)
While one can specify the choice of the functions ρ(s) and G(.) under which (2)
will be definitely true, we prefer to assume it explicitly in order not to overweight
the exposition. Indeed, these assumptions look as very natural ones. As the
manager’s share of the firm assets grows it becomes harder to fire him off and
30
investment in barter has more marginal effect on the entrenchment.
Let us now determine the first period choice of m and b.
At t=0 I chooses between investment in new product lines and barter to maximize
p ( s, b)U ( w I + s (m + βb)) s. to m+b=1
Clearly, it is equivalent to the problem
π ( s, b) + w I + s (1 − (1 − β )b) → max , b∈[0,1]
(3)
Here π(s,b)=ln(p(s,b)). In interior solution the choice of b is given by
π b′ ( s, b* ) = s(1 − β ) (4)
From (4) it follows that
*′
′′
bs = ((1 − b) − π sb′′ ) / π bb
(5)
which is less than zero under the assumptions (2) provided that S.O.C. for (3)
holds. Thus, as the firm’s share s the incumbent manager owns grows and manager
becomes more powerful; the need for barter is obviated since more direct ways of
personal aggrandizement appear.
3.4 Empirical results
Data description
The model in Section 5 predicts a negative relationship between manager’s share in
firm’s assets and share of barter in sales. For testing the above model we use
Goskomstat data and a dataset comprising 211 Russian manufacturing stock
companies, which are listed in a stock exchange. This dataset was created based
upon the information collected in the framework of the Information Disclosure
Program carried out by Federal Commission on Securities Market (FCSM). The
data consisted of the mandatory companies’ quarterly reports to FCSM for the year
31
19999. In order to fill the gaps in accounting balances and to calculate industry
means we use the Goskomstat data. However, despite the fact that these two data
sources complement each other, the data are far from being complete; this refers
especially to accounting data. That is why the number of observations for the
regressions presented below is less than 211.
Unfortunately, we do not have data that simultaneously include both barter and
ownership structure. So we have to use a substitute for them. As a proxy of barter
we take firm’s total revenues (line 210 in the balance sheet) minus total money
revenues (line 430 in the balance sheet) in 1998. Then, we divide it on annual
firm’s sales. Certainly, this proxy includes all non-monetary revenues i.e. barter in
a broad sense. Although this variable is intended to get an exact volume of barter
revenues the absence of strict rules and much ambiguity in Russian audit system
distort to a certain degree the real data. Nevertheless, such approximation seems to
be relevant as it follows from the table 4.1, where we regress barter data for year
1998 from dataset Barter in Russian industrial firms on the constructed proxy for
barter.
Table 4.1.
Barter (dataset Barter in
Russian industrial firms)
Barter (proxy)
.28
(.03)
Constant
.28
(.03)
9
N
288
R2
0.19
These data are available at the official FCSM Internet homepage.
32
To control for other explanations of barter we also include in our regressions firms'
size, regional and industry dummies constructed in a similar way as it is described
in section 4.
We use the following set of variables describing characteristics of ownership
structure:
! share_s
- percentage of equity belonging to the state
! share_m - percentage of equity belonging to the incumbent management team
! share_out -percentage of equity belonging to outside blockholders , measured
as total share of blockholders (shareholders owning more than 5% equity each)
minus share of blockholders belonging to the incumbent management team
minus share of the state if it exceeds 5%;
! num_block - number of blockholders
The values of all ownership variables belong to the first quarter of 1999; if these
data were unavailable, we use the values for the second quarter of 1999.
Although the model takes into account only percentage of equity belonging to the
incumbent management team we use also other variables to control for other
explanations of barter.
The summary statistics and pairwise correlations are provided in the Tables 4.2 and
4.3 below.
Table 4.2
Variable
Mean
Barter
Share_m
Share_s
Share_out
Num_block
0.63
3.85
4.55
31.57
1.86
Std.
Dev.
0.30
11.42
11.97
32.33
2.02
Min
Max
0.01
0
0
0
0
0.99
78
69
100
8
33
Table 4.3
Barter
Share_s
Share_m
Share_out
Num_block
Barter Share_s Share_m Share_out
Num_block
1.00
0.04
1.00
0.04
0.04
1.00
- 0.10
-0.11
0.04
1.00
-0.21
0.12
-.02
0.71
1.00
Empirical results
We run two OLS regressions for barter. In the first regression we use num_block as
a control variable; in the second we use Share_out. The results of the estimations
are shown in the Tables 4.4. and 4.5.
Table 4.4
Table 4.5.
Barter
Coef.
Share_m
Share_s
Num_block
Size
Msk
Ural
Asia
ind1
ind2
ind3
ind4
ind5
ind6
ind8
ind9
ind10
ind11
Constant
-.0004
.0010
-.02**
.007
-.24***
.01
-.05
.10
.02
-.05
-.23
.06
-.02
.13
-.12
-.32**
-.38***
.81***
N
R2
152
0.48
Std.
Err.
.0017
.0017
.010
.01
.09
.06
.06
.15
.15
.16
.17
.15
.14
.21
.16
.15
.14
.13
Barter
Coef.
Share_m
Share_s
Shareout
Size
Msk
Ural
Asia
ind1
ind2
ind3
ind4
ind5
ind6
ind8
ind9
ind10
ind11
Constant
-.0004
.0007
.00
.006
-.26***
.00
-.09
.06
-.01
-.10
-.25
.01
-.03
.14
-.11
-.35**
-.40***
.72***
N
R2
138
0.46
*** - significance at 1% level, ** - significance at 5% level, * - significance at 10% level
34
Std.
Err.
.0017
.0018
.01
.09
.06
.06
.15
.15
.16
.18
.15
.14
.22
.16
.15
.14
.13
It is seen that coefficients at share_m in both regressions, though have the predicted
sign, are insignificant. The same is true for the state share. The only parameter of
ownership structure that is significant is the number of blockholders. This result,
however, is not robust – as we substitute the number of blockholders by their share
we obtain the opposite result. The coefficient changes its sign and becomes
insignificant.
Overall, we see that coefficient at share_m is insignificant and this result is robust.
4. Conclusions
In this paper, we address two questions. First, we study the relationship between
debt and barter. Under the lack of effective bankruptcy procedures, barter can be
used to hide revenues from a firm's outside creditors. If the bankruptcy legislation
were strictly enforced, the firm that does not pay its debt would lose the control
over its assets to the creditors. If there is no threat of losing control, however, the
manager wants to sell for barter whenever he expects that cash will be taken by the
creditor and there will no other source of finance for purchasing inputs. We
explicitly model the possibility of re-financing the debt. It turns out, however, that
the firm's lack of commitment provides incentives for barter even if the firm and
the creditor can renegotiate the debt (since more cash implies weaker bargaining
position of the firm in the renegotiation). Another explanation of the high level of
barter is the presence of the government bonds (GKO) bubble in 1995-1998. If the
outside investment opportunity (GKO) yields a very high real interest rate, the
creditors will not be interested in refinancing the debt which would, of course,
provide incentives for using barter.
Our empirical analysis supports the predicted positive relationship between
indebtedness and barter. Which of the two explanations of this relationship (lack of
35
commitment or the GKO bubble) is correct? Both are consistent with
macroeconomic evidence. The period of the highest GKO yields coincided with the
highest levels of barter. Moreover, as the causality test in Brana and Maurel (1999)
shows, higher real interest rate caused higher barter in 1995-98. On the other hand,
the meltdown of August 1998 destroyed the GKO bubble bringing the real interest
rate to reasonably low levels. As one should have expected, barter has declined,
too. But it has not disappeared altogether and is still much higher than in other
economies.
Our model implies that hardening the bankruptcy procedures would have different
effect in the presence and in the absence of a bubble. If the bubble is present, lack
of bankruptcy leads to barter, which is definitely less efficient than money, but
helps to protect the firms’ working capital. Barter is a survival strategy that helps
firms with high indebtedness to survive and keep producing.
The returns to
production are lower than the GKO yield but since the latter is bubble, production
may be socially efficient. On the other hand if there is no bubble, lack of effective
bankruptcy procedures and creditor rights protections results in barter which is
simply less efficient than monetary exchange. Our analysis suggests two policies
that can decrease barter. First, avoid re-building high-yield debt financing of the
budget deficit. Second, continue the efforts in introduction of effective bankruptcy
procedures.
The other question we analyze is the possibility of using barter by Russian
managers in order to entrench themselves. We build a model in which we provide
some microeconomic foundations for considering barter in this role. The main
prediction of the model is negative relationship between share of barter in sales and
manager’s share of firm’s assets. Empirical analysis, however, rejects this
hypothesis. We may suppose two possible explanations. First is the bad quality of
36
data – quite a conventional argument pertaining to Russia. The other is that the
decision to use barter in Russia is determined by other, more rough forces like
those we survey in section 2 and consider in section 3 and 4.
References
Aukutzionek, Sergei. ''Industrial Barter in Russia'', Communist Economies and
Economic Transformation, 1998, 10(2), pp.179-188.
Brana, Sophie and Mathilde Maurel. ''Barter in Russia: Liquidity Shortage Versus
Lack of Restructuring”, CEPR Discussion Paper No. 2258, 1999.
Berglof, Erik and Ernst-Ludwig von Thadden. “The Changing Corporate
Governance Paradigm: Implications for Transition and Developing Countries”,
preliminary draft.
Commander, Simon and Christian Mumssen. ''Understanding Barter in Russia''.
EBRD Working Paper #37, 1998.
Edlin, Aaron S. and Stiglitz, Joseph E. “Discouraging Rivals: Managerial RentSeeking and Economic Inefficiencies”, American Economic Review, 1995, Vol.85,
N.5
Gaddy, Clifford and Barry W. Ickes. “To Restructure or Not To Restructure:
Informal Activities and Enterprise Behavior in Transition”, mimeo, PSU, 1998
Hart, Oliver. Firms, Contracts and Financial Structure. Oxford University Press,
1995.
Hayashi, Fumio and Akihiko Matsui. “A Model of Fiat Money and Barter”,
Journal of economic theory, Vol. LXVIII, 1996, 111-32
Hendlley, Kathryn, Barry W. Ickes and Randi Ryterman. “Remonetizing The
Russian Economy”, in H. G. Broadman, Ed., Russian Enterprise Reform: Policies
37
to Further the Transition, The World Bank, Washington, DC, November 1998.
Ickes, Barry W., Peter Murrell and Randi Ryterman, “End of the Tunnel? The
Effects of Financial Stabilization in Russia”, Post-Soviet Affairs, 13, 2, 1997,
pp.105-133.
Kiyotaki, Nobuhiro and Wright, Randel. “On Money as a Medium of Exchange”,
Journal of Political Economy, 1989, Vol. XCVIII, pp. 927-54.
Krugman, Paul (1988). "Financing vs. Forgiving a Debt Overhang", Journal of
Development Economics, Vol.29, 253-268.
LaPorta, Rafael, Florencio Lopez-de-Silanes, and Andrei Shleifer. “Corporate
Ownership Around the World”, mimeo, Harvard University, 1998.
LaPorta, Rafael, Florencio Lopez-de-Silanes, and Andrei Shleifer. ”Investor
Protection and Corporate Governance”,1999.
Myers, Steven and Raghuram Rajan. ''The Paradox of Liquidity''. Quarterly
Journal of Economics, 1998.
Polterovitch, Victor. “Institutional Traps and Economic Reforms”, NES Working
Paper, 1998.
Shleifer, Andrei and Vishny, Robert. W. “Management Entrenchment. The case of
Manager-Specific Investments”, Journal of Financial Economics, 1989, Vol. 24,
pp.123-139.
Shleifer, Andrei and Vishny, Robert. W. ”A Survey of Corporate Governance”, The
Journal of Finance, 1997, Vol. LII, N. 2.
Yakovlev, Andrei. “Classification criteria of non-monetary transactions in a
transition economy”, 1999a.
Yakovlev, Andrei. “Black Cash Tax Evasion in Russia: Its Form, Incentives and
Consequences at Firm Level”, mimeo, Higher School of Economics, 1999b
38
Документ
Категория
Типовые договоры
Просмотров
20
Размер файла
249 Кб
Теги
1/--страниц
Пожаловаться на содержимое документа