A Theory of Shortage in Socialist Economies Based on the "Soft Budget Constraint"

by Yingyi Qian
A Theory of Shortage in Socialist Economies Based on the "Soft Budget Constraint"
Yingyi Qian
The American Economic Review
Start Page: 
End Page: 
Select license: 
Select License

A Theory of Shortage in Socialist Economies Based on

the "Soft Budget Constraint"

This paper attributes shortages of goods in socialist economies to the soft financial constraints that firms in such economies face. A "soft budget constraint" problem arises when the state bank is unable to make a credible commitment not to refinance bad projects once some investment costs are sunk. In such a situation, if a consumer good is also demanded by firms as an input and the seller cannot separate firms from households, the high market-clearing price would lead to welfare losses because too many bad projects would start and crowd out household consumption. (JEL P20, D82, E10)

It has been long recognized that short- age is a persistent feature of all socialist economies. Consequences of shortage (misallocation of resources, delaying of completions of projects, queuing, hoarding, rent-seeking, etc.) are widely observed and extensively analyzed.' However, the ques- tion of why there is shortage in the first place has not yet been well understood. Economists are particularly puzzled by the fact that the government in a socialist econ- omy has immense power to fuc prices, at least in the long run. Given the costs of resource misallocation caused by rationing when prices are set too low, what prevents the government from raising prices to elimi- nate shortage altogether?

Explanations heard most frequently rely mainly on political considerations and, in

*Department of Economics, Stanford University, Stanford, CA 94305-6072. I owe special thanks to Janos Kornai and Eric Maskin, not only for their advice and encouragement, but also for their pioneering works on the subject. I also thank Carla Kriiger, Andreu Mas- Colell, Yijiang Wang, Martin Weitzman, Chenggang Xu, and participants in the 1990 SITE summer work- shop for comments and discussions, and three anony- mous referees for suggestions.

'~anos Kornai's (1980) book, Economics of Shortage, discusses in great detail shortage phenomena and eco- nomic and social consequences of shortage. In a recent paper, Martin L. Weitzman, (1991) analyzed deforma- tion of shortage in an element model.

particular, on alternative objectives of the state and its bureaucrats. At one end of the spectrum, some argue that low prices and rationing will ensure a more equal distribu- tion of goods to be allocated. This argument seems particularly relevant for the basic- need goods like food stuffs and housing (e.g., Weitzman, 1977). At the other end of the spectrum people argue that shortage benefits the middle- and lower-level state bureaucrats/managers by allowing them to maintain their power, to seek rents, and to have control over those who are rationed (e.g,. Kornai, 1992; Andrei Shleifer and Robert Vishny, 1992). There is no doubt that these (and other) political considera- tions provide some explanations for short- ages. However, there may be pure efficiency reasons, too. The purpose of this paper is to show that, even with a benevolent govern- ment that maximizes a distribution-free function of total social surplus, shortages may still persist under centralization of-fi- nancial resources. My approach here is complementary to, rather than a substitute for, other theories.

The new ideas of my theory are twofold. First, I link shortages of consumer goods to the demand of the sector of state-owned firms in view of the observation that the firm sector competes with the household sector for consumer goods to be used as inputs. Second, I relate the firms' demand to the financial constraints that state finan- cia1 institutions are able to impose and analyze the interactions between the real aspect and the financial aspect of the economy.

Specifically, my theory of shortage is based on the notion known as the "soft budget constraint," the problem that has been widely recognized as the fundamental one in all socialist ec~nomies.~

It is well known that in socialist economies state banks or other financial institutions are usually unable to avoid excessive renegotiation with state-owned firms in financial distress. Re- financing, subsidies, and bailouts of bad projects happen frequently, and unprofitable firms survive indefinitely. This in turn increases demand for investment ex ante. Compared to the case of the hard budget constraint in capitalist economies, demand for inputs from the firm sector is higher, other things being equal.

Competition between the firm sector, which has a soft budget constraint, and the household sector, which has a hard budget constraint, occurs when the same good is demanded by households for consumption and by firms as an input. Examples of such goods abound. Textiles, grain, gasoline, electricity, cars, housing space, and passen- ger and cargo transportation are all demanded by both households and firms, and all of them are in short supply in any social- ist economy. For those goods, I argue, if the seller cannot perfectly separate firms from households, a high market-clearing price will exclude many households from receiving consumer goods but will not deter ineffi- cient projects from starting because of the soft budget constraint. This clearly leads to social welfare losses.

However, at a lower price, social welfare

h he term "soft budget constraint" was coined by Kornai (1979). The soft-budget-constraint phenomena in socialist economies have been well documented by now in both the literature and the press (e.g., Kornai, 1986; Paul Bowles and Gordon White, 1989). Promi- nent economists actively involved in the transition programs all take the soft-budget-constraint problems seriously (e.g., David Lipton and Jeffrey Sachs, 1990; Ronald McKinnon, 1991).

may be higher. The idea is that, although a bad project may be able to secure the funds it needs under the soft budget constraint, it may not be able to obtain all the inputs it requires if they are in short supply. In the model I develop in this paper, a bad project is one that requires more time and capital for completion than its lender originally ex- pected. Hence, such a project will ultimately be refinanced in order to continue. Refi- nancing will do little good, however, if the project is unable to obtain sufficient inputs. Therefore, shortage may discourage a man- ager who knows his project is bad from applying for financing in the first place. This would improve the average quality of the projects. Of course, setting the price below the market-clearing level has inefficient consequences in terms of misallocation, pro- ject completion delays, and supply distor- tions. I show that the incentive effect may outweigh these adverse outcomes, however, in certain situations. In such a case, increas- ing the price to the market-clearing level is welfare-reducing, and therefore, a short- age results as an equilibrium phenomenon. In such an equilibrium, the price of an input/consumer-good acts as a screening device.

In contrast, the hard budget constraint of capitalist economies serves to screen out bad projects directly since such projects will find refinancing difficult. Therefore, under capitalism, there are no social benefits to keeping the prices of inputs/consumergoods below their market-clearing levels. That is, my model predicts that shortage is a phenomenon peculiar to socialist economies.

The notion of the soft budget constraint and its role as a cause of shortage was originally developed by Kornai (1979, 1980, 1986). Kornai argues that, because state-owned firms expect to receive sufficient fi- nancial support from the state to cover their expenditures (the soft-budget-constraint problem), their input demand is essentially inelastic. Therefore, there is in general no price that equalizes aggregate supply and demand, and shortage arises regardless of the price level. Although my idea of attributing shortage to the soft budget constraint is influenced by Kornai's work, my theory of shortage differs from Kornai's in an important respect. Kornai argues that there is no price which clears the market under the soft budget constraint. In his framework, price plays no allocative role, so one cannot consider costs and benefits asso- ciated with shortages. In my model, there exists a price that clears the market, but the state may not want to set the price at the market-clearing level in the presence of the soft budget constraint for efficiency rea- sons.

My model is built upon a recent work of Mathias Dewatripont and Eric Maskin (1990) in which the soft budget constraint is formulated in the context of sunk costs and contract renegotiation. The purpose of their paper is to show how centralization of credit leads to a soft budget constraint and how decentralization of credit may lead to a hard budget constraint. In contrast, the fo- cus of my paper is to explore the relation- ship between a soft budget constraint and shortage, that is, to explain shortage by tak- ing the soft budget constraint of firms as given.

The analvsis here is relevant to the recent discussions about the problem of transition from socialism to a market economy. Specifically, the issue of sequencing of price liberalization and reform of firm incentives has been a focus in policy debates for a long time.3 My model is an attempt to analyze formally the issue of price versus incentives, and my analysis suggests that financial disci- pline together with the corresponding insti- tutional changes is needed at the time of price liberalization. To the extent that the two reforms are complementary, price liber- alization without hardening the budget con- straint of the firms may induce adverse wel- fare consequence^.^ At a more general level, my analysis also suggests that the usual pol- icy recommendations drawn from standard price theory should be reconsidered before they are applied to socialist economies in

3~eeBarry Naughton (1992) on the Chinese debate on the subject in the 1980's. 4~heRussian experiencl in 1992 provided some evidence for this possibility.

which the soft-budget-constraint problem or other institutional problems are pervasive.

The paper is organized as follows. Section I introduces the model. Section I1 studies the incentive effect of shortages on firms and demonstrates that there are efficiency reasons the state may not want to set the price at the market-clearing level. Section I11 discusses extensions of the model, and Section IV compares my analysis with the literature.

I. The Model

The model has three sectors: the house- hold sector, the firm sector, and the finan- cial sector. In addition there is a price regu- lator. I consider the following two-good economy in order to study the problem of allocation of a good that is demanded by households for consumption and at the same time demanded by firms as a productive input. Good 1 is called money/output, which is the numeraire (i.e., the price of good 1 is normalized to 1). Good 2, the input/consumer-good, is demanded by households for consumption as well as by firms as inputs.

The economy has a stationary structure with infinitely many periods. In each period the households are endowed with some units of good 1(money income). The demand for good 2 (consumer good) by the household sector in each period is characterized by a demand function x(p), where p is the price of the consumer good. The inverse demand function v(x), where v(x(p))=p, represents the "willingness to pay" for each unit of the consumer good. The net consumer surplus in each period is given by

Firms have access to certain technologies (projects) that transform good 2 (input) into good 1 (output). In each period, N new projects are generated in the firm sector. The manager, an expected-utility maximizer, is risk-neutral. Firms have no initial endowments of either inputs or money, and managers must go to the state bank for credit in order to carry out the projects. To model the soft-budget-constraint problem in the financial sector I follow the recent work of Dewatripont and Maskin (1990).5.h

There are two types of projects: "good" and "bad." With one unit of input together with managerial effort e, a good project generates monetary return R, in the first period as well as nontransferable private benefits B, > 0 to the manager. The private benefits include increases in human capital, personal connections, power, status, pres- tige, and the like. With an initial one unit of input and managerial effort e, however, the bad project remains incomplete at the end of the period and so generates zero in that period. If the project is terminated at this time, the manager suffers private losses equal to 4 = $(el, which is the disutility of effort. However, with an additional one unit of input and additional effort e, the bad project can be completed in the next period, thereby generating R,, in which case the manager receives private benefits equal to B, > 0. Thus, bad projects are those that cannot be completed using only the initial outlay sufficient for a good pr~ject.~ To

%n Dewatripont and Maskin (1990), the soft budget constraint arises under centralization of credit. In such a case, because of sunk costs, refinancing projects that are ex ante unprofitable may become sequentially opti- mal once they have begun. Dewatripont and Maskin consider decentralization of credit as a commitment device against refinancing which can prevent unprof- itable projects from being undertaken. The idea is that if the original creditors are relatively small, refinancing must come from new creditors who may be forced to concede informational rents to their better-informed predecessors, which makes refinancing less attractive.

here are several other models developed to study the impact of the soft budget constraint on the behav- ior of state-owned firms (e.g., Kornai and Jorgen Weibull, 1983; Stephen M. Goldfeld and Richard E. Quandt, 1988; Mark E. Schaffer, 1989; Abhijit Banerjee, 1991). However, unlike Dewatripont and Maskin (1990), these models take the soft budget con- straint as exogenously given. Kornai's own explanation is mainly political: he emphasizes the "paternalistic" role of the socialist state.

hny feature that causes a project to make ineffi- cient use of its capital will render it "bad" in my sense. Because it is unprofitable initially to allocate to a project more than the "good" quantity of capital, the implication of this inefficiency is that bad projects will

simplify matters, I assume that, within each period, investment is made first, and then the applications of effort and input are made simultaneously.

The state bank, endowed with some units of money in each period, maximizes the expected returns. In addition to the above risky projects there is another safe invest- ment which guarantees a rate of return r. That is, the opportunity cost of capital to the bank is r. In order to focus exclusively on input rationing, I assume that the bank has enough money to finance all risky pro- jects so that there is no credit rationing. The bank approves (or rejects) a proposal if the expected net present value is greater (or less) than zero.

I assume that investments, returns, and managerial efforts are all observable. There are therefore no incentive problems regard- ing carrying out irlvestment, production, and repayment of money. The problem instead is how to turn away bad projects before they are started. To avoid trivialization of the problem I make the following assumption, under which the manager benefits from un- dertaking both types of project, even with- out any pecuniary rewards:



where 6 is the time discount ,factor.

Although investments, returns, and man- agerial efforts are all observable, the bank has imperfect information about the types

take longer to complete than good ones (more capital has to be allocated to the bad project after the initial allocation is exhausted). In fact, a common problem in socialist economies is that many projects do take longer to complete than expected. In both official documents and the press, the central planning commission constantly complains about too many unfinished projects that drive away funds from starting new projects.

of the projects. Managers know from the beginning whether their projects are good or bad, but the bank only obtains this infor- mation after the first unit of input is sunk. However, the bank does know that the pro- ject is "good" with probability a and is "bad" with probability 1-a. The bank therefore faces an adverse-selection prob- lem. Without an additional screening mech- anism, the bank either approves all projects or none of them.

In my model, when the type of project is revealed after the initial investment is sunk, the bank can decide to terminate bad pro- jects. Indeed, in a dynamic setting such as mine, an option to terminate a project after the initial phase gives banks an opportunity to punish managers running bad projects. When this threat is credible, banks can ef- fectively deter credit applications from bad projects and therefore prevent them from starting. Such an incentive mechanism is absent in any static model.8

However, the dynamic setting creates a problem which is also absent in a static model: renegotiation. The key to the model I present here is that I allow renegotiation between the bank and the firm after the type of project is revealed. Because the first-period costs are sunk, termination of bad projects (and therefore punishment of the managers) is not "renegotiation-proof." I give all bargaining power to the bank, so that the bank is able to extract all the mon- etary returns from financing the projects, and the managers are only limited to non- transferable private benefits. Since -4 < 0 but B, > $, managers with bad projects al- ways have the incentive to continue their projects. Moreover, the bank is also better off in refinancing the additional one unit of input if R, > p, since the first-period costs are already sunk. When R, > p, termina- tion of bad projects is not ex post optimal, so it is not ex ante credible; but allowing

or example, in the static model of Joseph E. Stiglitz and Andrew Weiss (1981), creditors cannot deter bad projects.

bad projects to continue creates inefficiency ex ante.

Finally, the supply s(p) of the consumer- good/input in each period is determined by a cost function 4s) such that p = C'(S(~))." I consider the stationary situation in which the price p remains constant from period to period. The role of the price regulator, the representative of the benevolent state, is to control the price p to maximize (per-period) total social surplus, which includes consumer surplus, the bank's profits, and the profits from the seller of the consumergood/input.10

11. Shortage and the Soft Budget Constraint: The Incentive Effect

In the presence of a soft budget constraint in the credit market, either all pro- jects are financed, or none of them are. The interesting case is the former. If all projects are financed and there is no shortage of inputs, N new projects start in each period. Of the N new projects, only aN good pro- jects are completed in the current period and (1-a)N bad projects continue through to the next period." In addition, there are (1-a)N bad projects which started in the last period but finish in the current period.

Let p* be the market-clearing price; that is, p* satisfies

x(p*) + ~[cu+2(1-a)] = s(pX)

Notice that demand from the bad projects which have a soft budget constraint competes with demand from the households who have a hard budget constraint. Thus, social welfare suffers from the soft budget constraint.

'~ote that in this "upstream" industry, input is good 1 (money/output), and output is good 2 (consumer good/input).

''1 exclude managers' rents from the total social surplus for simplicity, which should not change the results qualitatively.

11The underlying assumption is that N is large, and therefore the sample distribution of types of projects is the same as the population distribution.

Clearly, if the bad projects can be distin- guished by lenders ex ante from good pro- jects, social surplus is maximized at the market-clearing price. Next, I examine the welfare implications of pricing when infor- mation about projects is imperfect.

Case 1: p 2p*.-When the price is set above p*, goods are always available as long as one has money to buy. A good project is expected to be completed in one period and a bad project is expected to be completed in two periods. Given Assumption 1, there is no reason for managers with either type of project not to apply for credit. If I denote R', and Rb as the expected net present value of a good and a bad project, respec- tively, R',=Rg-p and Rb=R,/(l+r)- p -p/(l + r). The bank approves all pro- posals if aRI, +(l- a)R', > 0. In such a case, maximizing the total social surplus in each period is equivalent to maximizing

where N(1- a)rp is the opportunity cost of capital.

PROPOSITION 1: The price regulator never wants to increase the price of the input/ consumer-good aboue the market-clearing leuel p* .I2

PROOF: At any p2p*,

which is nonpositive because the price is not

12~iven that consumers cannot be distinguished from firms and that good firms cannot be distinguished from bad firms, the only policy tool at the state's disposal is price (I abstract from the issue of alterna- tive rationing schemes, since this is a second-order consideration).

lower than the marginal cost for p 2 p*.

Case 2: p <p*.-If the price p is set below p* ,aggregate demand exceeds aggre- gate supply, and the input/consumer-good is rationed. For simplicity, I assume the random-rationing rule in my model. Let

q =(aggregate supply)/(aggregate demand)

for p s p*. Then everyone-a consumer, a firm with a good project, or a firm with bad project-is able to buy all he wants with the probability q but gets nothing with probabil- ity 1-q.13 When consumers do not get con- sumer goods, they have to assume the other good ("money"), and they consequently lose consumer surplus. When firms do not receive inputs, they have to wait until the next period, and their projects are delayed. I assume that money is granted at the instant credit is approved, and the bank cannot take the money back if it is unspent due to shortage.

When there is a shortage of inputs the project may not be completed even though the firm is granted credit. Since on average a firm needs l/q periods to obtain inputs, the average time for a good project to be completed is also l/q. Let R',(q) represent the expected net present value of a good project if the firm is able to receive inputs with probability q. Because the probability that the project is completed and the bank is paid at the end of the period is q and the probability that the project is delayed to the next period is 1-q, the following basic re- cursive equation is obtained:

Rh(q) = q(Rg-P)+(1-q)[~',(q)/(l+ r)]

from which I derive

13~herandom-rationing scheme can be supported by (i) issuing coupons to all households and firms with some lucky numbers (lottery) or (ii) a random queue wherein one's position in a line has a uniform probabil- ity distribution. One buys all he wants if he holds the winning number of the lottery in the former case, and if the good is still available when he moves to the first of the line in the second. The same rationing scheme is also used by Weitzman (1991).

which implies

Therefore, R',(q) increases in q; that is, shortage always reduces the bank's net re- turn. This is because shortage delavs the u, =q(Bg -$)/[I -s(1- q)] .

completion of the project, which is-costly when the opportunity cost of capital, r, is greater than zero.

Let Rb(q) be the net present value of a bad project and let R',(q) be the net present value of a bad project after the first unit of investment is sunk, assuming that the prob- ability of receiving inputs is q. similar to K',(q), then

Therefore, the bank will refinance the bad project after the first phase is completed if and only if R, > p, the decision not affected by the presence of shortage. If the bad project is refinanced, the following recursive equation should hold:

which implies

Shortage also reduces the bank's net return from a bad project because it needs on average 2/q periods to be completed.

PROPOSITION 2: The bank's decision on refinancing bad projects does not depend on shortage in the input market. The bank ap- proves all projects or denies all projects according to whether aR',(q) +(1+ cr)Rb(q) is positice or negative.

I now turn to the manager's incentives of undertaking projects under shortage. Let U, be the discounted net benefit to a manager who undertakes a good project. A similar recursive equation is obtained:

Because U, increases in q when 6 < 1, shortage reduces the manager's net benefit when he is impatient. However, shortage does not affect his incentive to apply for credit since he always gains from undertak- ing a good project:

PROPOSITION 3: The manager with a good project always wants to apply for credit, regardless of shortages of inputs.

Let U, be the discounted net benefit to the manager who undertakes a bad project, assuming that the bad project is refinanced. Also let U, be the discounted net benefit to the manager with a bad project after the first-period investment is carried out. Then,


Although shortage does not change the manager's incentive for obtaining refinanc- ing ex post as U, > 0, shortage may change the manager's incentive to apply for credit ex ante if his project is bad.

PROPOSITION 4: The nzanager with a bad project applies for credit if and only if B, > [2+(1/6 -l)/ql*.

Under the soft budget constraint, managers are not afraid of starting bad projects because they know that the bank is unable to commit for not refinancing. According to Assumption 1, B, >(1+ 1/6)$, managers certainly have incentives to apply for credit when there is no shortage (i.e., q = 1). The presence of shortage increases private costs to the manager since [2+ (1/6 -l)/q]$ > (1 + 1/6)$ for 6 < 1 and therefore reduces his net benefit given private benefits B, of undertaking a bad project. When q becomes sufficiently small, that is, when short- age is sufficiently severe, the manager with a bad project will not apply for credit in the first place.

The intuition behind this result is the following. Under the soft budget constraint, although managers are sure that bad pro- jects will always be refinanced, they are not sure they will be able to acquire the inputs in short supply. Of course, managers with good projects are not sure they will be able to get inputs either. Nevertheless, in my model, a bad project lasts longer than it should because of excess renegotiation and refinancing, the problem of the soft budget constraint. Hence, the presence of shortage is more costly to the manager with a bad project than to the manger with a good project.

The manager with a bad project is less likely to apply for credit if shortage is in- tense (i.e., q is small), the manager is less patient (i.e., 6 is small), his disutility is high, or the benefits are low. Imagine that the population of managers is characterized by different B,'s, a's, and 4's. Let A(q) be the percentage of managers with bad projects applying for credit when the probability of receiving input is q. For all 6 < 1, A(q) is a nondecreasing function of q with A(0) = 0 and A(1) = 1. Because the expected time to complete a good project is l/q and the corresponding time for a bad project is 2/q, the average total number of projects at any time is

Therefore, the input allocated to the firm sector under the random-rationing rule is N[a +2(1- a)~(~)].'~

141f all managers are identical, shortage q such that R, =[2+(1/6 -l)/q] will make managers with a bad project indifferent between undertaking and not under- taking projects. Then the equilibrium percentage of bad projects A will be determined by the supply and demand: qx(p)+ N[a +2(1- a)A] = s(p).

PROPOSITION 5: The percentage of bad projects undertaken is a nondecreasing func-

tion of p for p Ip*.


For any given p <p*, function qx(p) + aN +20-a)NA(q)- s(p) is a strictly in- creasing function of q and is greater than 0 for q = 1. Therefore, there exists a unique q(p) such that

Furthermore, q(p) is an increasing function of p with q(p*) = 1. Hence A is a nonde- creasing function of p for p Ip*, and A is equal to 1 for p =p*.

Shortage reduces the incentives of man- agers with bad projects to apply for credit and therefore improves the overall quality of the projects that are financed. In this sense, the price of the input/consumer-good is used as a screening device for project selection in the credit market.

One implication of this theory is that relaxation of price controls may be accom- panied by a worsening of the soft-budget- constraint problems, which in fact has been observed during China's reforms. Before the economic reforms, almost all prices in China were controlled by the state. A series of steps has been undertaken by the Chinese government since 1979 to relax price con- trols. As a result, prices now fall into three categories according to the degree of state intervention: (i) planned prices, (ii) guided prices, and (iii) market prices. By 1990, ac- cording to official statistics, the latter two categories accounted for more than 70 per- cent of total retail value. Furthermore, 52 percent of the value of agricultural products and 37 percent of the value of wholesale industrial goods corresponded to prices set by the market alone (People's Daily, overseas edition, 29 November 1991, p. 1).

Notwithstanding these reforms, the soft- budget-constraint problem seems to have worsened. Specifically, in the past few years, there has been a rapid increase in the num- ber of loss-making firms as price controls have been relaxed. Losses of industrial firms with negative profits increased from 3.42

billion yuan in 1984, the year major urban reforms were started, to 10.66 billion yuan in 1988, and further to 45.37 billion yuan in 1990 (Statistic Yearbook of the Chinese In- dustrial Economy 1991). The number of loss-making state-owned industrial firms in- creased from less than 10 percent of the total in 1985 to about one-third in recent years.15 This positive correlation between price liberalization and increases in the number of loss-making firms is consistent with Proposition 5: as price controls are relaxed, managers are encouraged to under- take more bad projects since inputs are easier to obtain.16

Next, I will analyze welfare implications of pricing policies. In this model there are three types of distortion when the price is set below the equilibrium level p*: first, goods may not be assigned to those whose valuation is the highest; second, completion of the projects is delayed; and third, supply of the input/consumer-good is distorted. However, the incentive effect of shortage is beneficial under the soft budget constraint. For p 5 p* the total social surplus is given by

151t should be noted that the above figures are expressed in current prices. Between 1984 and 1990, the national retail price index in China rose by 76 percent (Statistical Yearbook of China 1991). Therefore, there was indeed a significant real increase in losses. Other evidence on the worsening of the soft- budget-constraint problem during China's reforms is presented in Bowles and White (1989).

16several other explanations are also possible. For example, accounting losses may have resulted from growing evasion due to the loosening of central control during that period. Another reason for the sharp in- crease of losses between 1989 and 1991 was the govern- ment's austerity program and the economic recession that followed.

where rpN[a(l-q)+ (2 -q)(l- a)A(q)] is the opportunity cost of capital.

PROPOSITION 6: A marginal decrease of the price from p* may lead to an increase in the total social surplus. In particular, if A(q) is differentiable at q = 1 and r = 0, then a marginal decrease of the price from p* leads to an increase in the total social welfare if and only if

PROOF: For all p p*, differentiating both sides of the identity


At p = p*, q = 1 and the left derivative of W with respect to p is

Hence, dW/dp < 0 if the term in braces is negative and qt(p*)> 0. When r = 0, the

left derivative dW/dp < 0 at p = p* if and only if inequality (1) holds.

The left-hand side of inequality (1) is the marginal cost of the misallocation effect due to random rationing. Because the cost of delay disappears when r = 0 and the distor- tion of supply is second-order around p = p*, the left-hand side of (1) represents the total marginal cost of distortion due to devi- ating from p*. On the other hand, the right-hand side is the marginal savings from the incentive effect of reducing the percent- age of bad projects. Inequality (1) shows that lowering price is likely to increase the total social surplus if (i) there are initially more bad projects and fewer good projects (i.e., 1-a is large); or (ii) losses from the bad projects are large (i.e., 2p* -R, is large); or (iii) the incentive effect of short- age on bad projects is strong (i.e., A'(1) is large); or (iv) the misallocation effect on households is small. Consequently, under these circumstances shortage is more likely.

In my theory, the reason for the govern- ment not raising prices to the market-clear- ing level under the'soft budget constraint is the fear of inviting more bad projects to start and therefore crowding out demand from the households. This is because in the presence of the soft budget constraint, the credit market is unable to screen out the bad projects, and the adverse-selection problem regarding the project quality is car- ried over to the input/consumer-good market. Through the incentive effect of short- age, the price of the good is used as a screening device for project selection. In contrast, under the hard budget constraint in capitalist economies, efficient project se- lection is made in the credit market, so aggregate demand includes only demands from households and good projects. The price plays the usual role of resource alloca- tion; there are no benefits in regulating the price below its market-clearing level.

111. Extensions: Partial and Complete Separation of Households and Firms

My analysis so far has assumed no sepa- ration of households and firms as far as the allocation of good 2 is concerned. This as- sumption is made in order to underscore the effect of com~etition between the firm sector and the household sector for the good to be used for consumption and for productive inputs. In this section I will dis- cuss briefly the situation in which house- holds and firms can be separated to some extent.

In my model the economy will be more efficient if the state is able to control the allocation of good 2 to households as a consumer good and to firms as an input and to charge differential prices in the two sec- tors. Indeed, socialist governments have tried hard to senarate firms from house- holds." However, because of the consider- able monitoring required, separation has never been complete; there have always been considerable flows between the two sectors.18

Consider the situation in which a propor- tion p (0 < p I1) of managers is able to obtain good 2 (inputs) from the household sector. Suppose that the government allo- cates sf units of good 2 to the firm sector and s, units to the household sector (for simplicity, assume sf + s, = o is a constant). Let p(s,) be such that

That is, p(s,) is the market-clearing price if s, units of good 2 are allocated to the household sector. I will show that the fol-

or example, sometimes firms have been required to buy inputs with "institutional checks" rather than cash.

his is so not just because of the existence of secondary or black markets. (Indeed, if such markets worked very well in practice, shortages would probably not have occurred; and, because such markets work against the screening role of shortages, a benevolent state in my model would try to prevent them from arising.) Rather, it is all too easy for a member of a firm to pose as a consumer if consumer prices are more favorable.


lowing condition is sufficient to produce equilibrium shortage in the household sec- tor even if the state can choose sf and s, optimally:

For any choice of sf > 0 and s, < w the market-clearing price in the household sec- tor p(s,) must be higher than p(w). Be- cause the left-hand side of the above expression is a decreasing function of p, the inequality must also hold for any p(s,). Therefore, following the proof of Proposi- tion 6, total social welfare will be improved by marginally lowering the household price from the equilibrium level p(s,), regardless of which s, is chosen by the state.

In the extreme case of complete separa- tion of households and firms (i.e., P = O), consumer-goods shortages within the house- hold sector will be completely eliminated in my model. This is because the government sets the price for the households only to maximize the consumer surplus, which must be the market-clearing price under the hard budget constraint. However, even under this situation, my arguments suggest that input shortages in the firm sector may still persist. According to Proposition 5, the incentive effect of shortage may improve the average quality of projects undertaken, independent of the presence of the household sector. Therefore, even when complete separation is possible, my theory still explains short- ages in the firm sector and therefore remains relevant to the policy debate about price liberalization.

IV. Concluding Remarks

Using incentive theory along with a given information structure to explain certain (Walrasian) "disequilibrium" phenomena is common to my theory of shortage in social- ist economies, as well as the "efficiency wage" theory of unemployment (e.g., Carl Shapiro and Stiglitz, 1984) and the theory of credit-rationing based on limited liability af- ter bankruptcy (e.g., Stiglitz and Weiss, 1981) in capitalist economies. The key to the last two theories is that, when there exists imperfect information, either the price (e.g., the interest rate) is used as a screening device or disequilibrium itself (e.g., unem- ployment) is used as a discipline device. Similar incentive and screening mechanisms are at work in my model as well.

However, it is worth noting the differ- ences. Central to the theories of unemploy- ment and credit-rationing is the question of why the competitive force of markets does not make wages fall or interest rates rise to clear the market. In my model, the question is why the benevolent government does not raise prices of consumer goods to eliminate shortage. Equally important, these theories essentially focus on only one market at a time: in efficiency-wage theory, it is only imperfect information in the labor market that is relevant, while in the credit-rationing theory, it is only imperfect information about projects that is relevant. In my theory of shortage of input/consumer-goods, imperfect information (adverse selection) is pre- sent initially in the credit market. It is then carried over to the product (input/ consumer-goods) market under the soft budget constraint. The interaction of short- ages in the input/consumer-goods market and project selection in the credit market is the focus of my analysis.

My analysis is also related to the recent literature on macroeconomic consequences of financial constraints in market economies. For example, Ben Bernanke and Mark Gertler (1989) analyze the relationship be- tween the net worth of firms and business fluctuations; Bruce Greenwald et al. (1990) study productivity growth of firms with the possibility of equity constraints. In this liter- ature, the agency cost associated with im- perfect information imposes constraints on the borrowing capacity of privately owned firms, which in turn has important macroe- conomic implications. In my model the agency problem, together with the inability of commitment of the state bank, leads to


too much borrowing of state-owned firms, which is the cause of shortages.


Banerjee, Abhijit. "Regulation and the Soft Budget Constraint." Mimeo, Princeton University, 1991.

Bernanke, Ben and Gertler, Mark. "Agency Cost, Net Worth, and Business Fluctua- tions." American Economic Reuiew, March 1989, 79(1), pp. 14-31.

Bowles, Paul and White, Gordon. "Contradic- tions in China's Financial Reforms: The Relationship between Banks and Enter- prises." Cambridge Journal of Economics, December 1989, 13(4), pp. 481-95.

Dewatripont, Mathias and Maskin, Eric. "Credit and Efficiency in Centralized and Decen- tralized Economies." Mimeo, Harvard University, 1990.

Goldfeld, Stephen M. and Quandt, Richard E. "Budget Constraints, Bailouts, and the Firm under Central Planning." Journal of Comparatice Economies, December 1988, 12(4), pp. 502-20.

Greenwald, Bruce; Salinger, Michael and Stiglitz, Joseph E. "Imperfect Capital Market and Productivity Growth." Mimeo, Stanford University, 1990.

Kornai, Janos. "Demand versus Resource Constrained Systems." Econometrica, July 1979, 47(4), pp. 801-19. .Economics of shortage. Amsterdam: North-Holland, 1980. . "The Soft Budget Constraint." Kyklos, 1986, 3901, pp. 3-30. . The socialist system. Princeton, NJ: Princeton University Press, 1992.

Kornai, Janos and Weibull, Jorgen. "Paterna- lism, Buyers' and Seller's Market." Mathematical social sciences, November 1983, 6(2), pp. 153-69.

Lipton, David and Sachs, Jeffrey. "Creating a Market Economy in Eastern Europe: The Case of Poland." Brookings Papers on Economic Actiuity, 1990, (I), pp. 75-147.

McKinnon, Ronald. The order of economic liberalization: Financial control in the tran- sition to market economy. Baltimore, MD: Johns Hopkins University Press, 1991.

Naughton, Barry. "Growing Out of the Plan: Chinese Economic Reform, 1978-90." Mimeo, University of California-San Diego, 1992.

Schaffer, Mark E. "The Credible-Commitment Problem in the Center-Enterprise Relationship." Journal of Comparative Economics, September 1989, 13(3), pp. 359-82.

Shapiro, Carl and Stiglitz, Joseph E. "Equi- librium Unemployment as a Worker Dis- cipline Device." American Economic Re- uiew, June 1984, 74(3), pp. 433-44.

Shleifer, Andrei and Vishny, Robert. "Pervasive Shortages under Socialism." Rand Jour- nal of Economics, Summer 1992, 23(2), pp. 237-46.

Statistical Yearbook of China 1991 (Chinese Edi- tion). Beijing: China Statistics Press, 1992.

Statistic Yearbook of the Chinese Industrial Econ- omy 1991 (Chinese Edition). Beijing: China Statistics Press, 1992. Stiglitz, Joseph, E. and Weiss, Andrew. "Credit Rationing in Markets with Imperfect In- formation." American Economic Reuiew, June 1981, 71(3), pp. 393-410.

Weitzman, Martin L. "Is the Price System or Rationing More Effective in Getting a Commodity to Those Who Need It Most?" Bell Journal of Economics, Autumn 1977, 8(2), pp. 517-24. ."Price Distortion and Shortage De- formation, or What Happened to the Soap?" American Economic Reciew, June 1991, 81(3), pp. 401-14.

  • Recommend Us