OCTOBER

1998

A SIMPLE FOUR SECTOR MODEL OF

RUSSIA’S “VIRTUAL” ECONOMY1

Clifford Gaddy Barry W. Ickes

The Brookings Institution The Pennsylvania State University

1. INTRODUCTION

Extensive use of barter and other noncash modes of payment, massive tax and wage arrears,
growing mutual indebtedness in the enterprise sector—all of this is by now so familiar a part of the Russian
economy that it is inescapable. The scale of demonetization in Russia today may be unprecedented in an
industrial society. In 1997, more than 50% of all sales by industrial enterprises in Russia were
nonmonetary.2 Over 40% of all taxes paid to the Russian federal government in 1997 were in made in
nonmonetary form.3 The rate of demonetization of local and regional budgets is even higher.4 Nonpayment
of wages and salaries is again on the rise. As of January 1, 1998, the average industrial worker was owed
nearly two months’ of back wages; in agriculture the average delay was 4.1 months.5

1
We are grateful to seminar participants at Brookings, Columbia, and the National Bank of Finland, and
especially to Rick Ericson, Ed Dolan, Gregory Grossman, Gur Ofer, John Steinbrunner, and Pekka Sutela for
helpful comments and discussion.
2
See Hendley, et al.,(1988) for a discussion of barter in the Russian economy. When the focus is narrowed to
the largest enterprises this ratio rises to 73%. [Karpov percentage of nonmonetary sales was 73% for the 210
largest enterprises. REB figure for barter was only about 50%, but REB sample does not include enterprises over
2,000 employees—i.e., the REB and Karpov samples are probably mutually exclusive.]
3
[From my data on the budget. I have to get the actual cites from earlier documents, since we excluded the
original footnotes from later versions of papers. In 1995 and 1996, 24% of all federal taxes were paid in non-
monetary form; in 1997 the non-monetary share was 41%.]
4
For a sample of 39 regions (of Russia’s total 89), the average share of nonmonetary tax revenues in 1996 was
60% to regional (oblast, kray, and republican) budgets and 43% to local (district and city) budgets. OECD
Economic Surveys, Russian Federation 1997, Paris, 1997, p. 181.
5
[Need source.]
FOUR SECTOR MODEL Page 2

The conventional wisdom seems to be that this is a situation out of control. Inter-enterprise debt and
tax arrears, one typical description goes, have combined to form a “vicious circle of bad debt, heavy
borrowing and high interest rates which is choking off economic growth.”6 The implication of the article
from which this quote is taken, and others like it, is that the barter-and-arrears economy is a morass that
sucks in innocent victims and ultimately threatens their ruin, or at least holds them back from achieving
economic success.

The purpose of this paper is to argue that just the opposite is the case. We believe that economic
performance in Russia is clouded by the existence of a virtual economy. The virtual economy is the result of
a pretense that the Russian economy is much larger than it really is. The pretense that more value is
produced than is actually the case is the cause of many of the puzzling phenomena in Russia today. The
virtual economy, we believe, is a fairly well-organized system. Rather than being victims, its
participants—which include not only enterprises, but also workers and governments—are largely willing
players. And rather than representing a threat to them, it is the virtual economy that is their only hope for
survival.7

To make our case, we present a very simple model of the virtual economy that we believe captures
the essence of what is happening in Russia today. In a companion paper to this one, we analyzed the
microeconomics of the typical Russian enterprise in the transition, as it chooses whether or not to
restructure. There, we emphasized the role of the enterprise’s endowment at the beginning of transition, not
only of physical and human capital, but also of “relational capital,” or contacts and connections with
government officials and other enterprise directors. We showed that the possibility of using relational capital
allows enterprises to survive without restructuring.8

In this paper, we step back and examine how the enterprise making this decision fits into the bigger
economic system that characterizes the virtual economy. We believe that the stylized model we present
below not only captures some of the most salient and unique features of the contemporary Russian economy,
but it also provides a convenient framework to discuss the implications for two of the other sectors, in
addition to the enterprise sector, that make up the virtual economy: the household sector and the government
sector.9

6
This phrase was taken from an AFP release dated 30 April 1998, by Jon Boyle, “Reformers tighten grip on
economy as Chubais takes over at UES.”
7
Our use of the term “virtual economy” is an extension of the language used in the December 1997 report of
the so-called Karpov Commission. Set up in 1996 by the Russian government to study the problem of nonpayments
and arrears in the economy, the body’s official title was the Inter-Agency Balance-Sheet Commission, under the
chairmanship of P. A. Karpov. The report concluded, among other things, that barter, by its use of “nonmarket or
‘virtual’ prices, [created] illusory or ‘virtual’ revenues, which in turn lead to unpaid or ‘virtual’, fiscal
obligations.” In Gaddy and Ickes (1998), we suggested that the appropriate name for the sector of the current
Russian economy that operates in this nonmonetary realm is “virtual economy.”
8
Barry W. Ickes and Clifford G. Gaddy, “To Restructure or Not to Restructure: Informal Activities and
Enterprise Behavior in Transition,” draft, May 1998.
9
Our paper is purposefully simplified, and thus we make strong assumptions to clarify the argument. For an
interesting analysis of some of the details of how the virtual economy operates, see Thompson (1998).
FOUR SECTOR MODEL Page 3

2. ROOTS OF THE VIRTUAL ECONOMY

The roots of the virtual economy lie in the largely unreformed industrial sector inherited from the
Soviet period. At the heart of the phenomenon are the large number of enterprises that still produce goods
but destroy value. An enterprise destroys value when the value of inputs purchased from other enterprises
exceeds the value of the output that is produced.10

To understand the phenomenon of value destruction it is important to begin with Soviet pricing.
Raw material inputs were underpriced in the Soviet economy. Their prices were based on the operating costs
of extraction, ignoring rent (that is, disregarding the opportunity cost of using the resources now rather than
in the future). No doubt this harmonized with the goal of increasing production today; scarcity pricing might
have induced more conservation, which mitigates against maximizing current production. This bias in raw
material prices fed into the system of industrial prices. Heavy consumers of energy were, in effect,
subsidized. So too were heavy users of capital, thanks to the absence of interest charges. In short, costs of
production were calculated on the basis of an incomplete enumeration of costs.

In addition to incomplete cost-based pricing, the system was biased towards certain users. The same
commodity would carry a different price if it were used by heavy industry or light industry. This would then
feed into the calculation of costs of production of these goods, so that high priority sectors would appear to
have lower costs of production than low priority sectors. This meant that the apparent distribution of
productivities at the onset of transition, what we may think are efficient sectors, was liable to mask the true
picture.11

The fact that the pricing system disguised the relative efficiency of various activities means that
only with economic liberalization would the true viability of these activities become apparent. Many sectors
that appeared to be creating value turned out to be destroying value once prices moved to reflect costs. The
extent to which the Soviet economy produced the “wrong things in the wrong way” could only be gauged
after liberalization. This effect was magnified by the move to world prices.12 Many industrial enterprises
could not cover costs once prices moved to market-clearing levels. Raising prices only led to unsold output.
Price liberalization revealed the extent to which value added in the Soviet economy was really created in the
energy and raw materials sector, but it had the effect of making reform appear to be the destroyer of the
manufacturing sector.

How could a value-destroying sector survive six years of market reform? The reasons are complex,
but the most important is that in Russia today enterprises can operate without paying their bills. This is
possible because value is redistributed to them from other sectors of the economy. One way this is done is
through tax arrears, which are in effect the continuation of budget subsidies in a different form. More
important, however, is direct redistribution of value to value-subtractors from the value-producing sectors of
the economy, primarily the resources sector.

It is important to understand the continuity with the past that is involved here. The Soviet economy
appeared to be a large industrial economy. In fact, industry in the Soviet economy was subsidized by under-

10
For a more detailed discussion of the concept of value subtraction, see the Appendix.
11
See [Ericson (1996)] for an analysis of the implications of arbitrary pricing on the apparent and actual
production of value added in the Soviet economy. [Ericson (1988)] was the first study to formalize the dual nature
of the Soviet economy in terms of priority (military) and nonpriority sectors.
12
This point was emphasized in Bosworth and Ofer (1995:80).
FOUR SECTOR MODEL Page 4

priced raw materials and insufficient charges for capital. The economy appeared to have a large
manufacturing sector that produced value; in fact, large parts of the manufacturing sector destroyed value,
but this was masked by arbitrary pricing.13 The roots of the virtual economy lie in the maintenance of this
pretense.

3. A SIMPLE ACCOUNTING MODEL

We analyze the operation of the virtual economy by studying a simple static accounting model. The
purpose is twofold. First, this model allows us to see how the parts of the economy interact. Second, the
model demonstrates how the virtual economy masks the true state of affairs.

The model is highly stylized. There are four sectors:

 H (“HOUSEHOLDS”)
 G (a value-adding industrial sector, “GAZPROM”)14
 M (a value-subtracting industrial sector, “MANUFACTURERS”)
 B (a government sector, “BUDGETS”)

The household sector (H) supplies labor and is the ultimate recipient of income. The government sector (B)
collects taxes and distributes the proceeds in transfers. We assume, for simplicity, that it transfers all tax
revenue to the household sector.15 There are two sectors of industry in our model: a resource-producing sector
(G), whose output is used as an input for a final goods producer (M).16

3.1 PRODUCTION

We assume that G and H each produce 100 rubles of output (gas and labor) at no cost. Gas is used as
an intermediate good in the production of M, which also uses the labor supplied by H. M’s product has a

13
It is perhaps more correct to say that the end-users in the Soviet regime -- the Communist Party -- placed
such a high value on the output of the defense sector that value was produced. The problem is that with the end of
the regime that value of that production has shrunk dramatically.
14
Strictly speaking Gazprom is not the only supplier of value into the system. Value is produced by other
energy producers (e.g., Lukoil), new enterprises, as well as those enterprises that have successfully restructured.
Gazprom is the most important example, and because of its role in supplying energy (via supplies to electricity
producers), it has the widest impact on the economy. It also means that value is pumped directly to lossmaking
enterprises, since they all consume energy.
Inside the Russian economy, Gazprom’s role is widely recognized. Consider the following quote by the
CEO of one of the major power companies in Russia, Mosenergo, Nestor Serebryannikov:

“The main income in the country comes from the resource-extracting industries, the production and sale
of gas and oil. Gazprom is the salvation of the government and the creditor of the rest of industry,
including the energy production sector.” [Ekonomika I zhizn', Moscow edition, no. 15 (April) 1998, p. 12.]
15
As we suggest in Gaddy and Ickes (1998), however, the government has an important role in the virtual
economy as the referee between competing interests. We ignore this is in the simple version of the model.
16
This is a stylized model. We are not asserting that value added in Russia is only produced in the resource
extraction sectors, nor do we assume that all manufacturing enterprises are value destroying.
FOUR SECTOR MODEL Page 5

market value of 100. In other words, M is a value-subtractor.17 We can illustrate the real flows in this

Figure 1
economy with a schematic diagram (Figure 1).

We assume that all agents in this economy attempt to meet their obligations. No one conceals income.
Thus, for example, M contracts to pay G and H the market value of the inputs (gas and labor) which they
supply. All accounting (between enterprises and for tax purposes) is on the accrual basis—i.e., income and
expense items are recognized as incurred, regardless of when they are actually paid.

The household sector requires a minimum of 100 in cash to survive. This is the only cash requirement
in the system. The output of M can be sold for cash at any time, at its market value.

3.2 FISCAL CONSIDERATIONS

We make the simplest possible assumption with respect to the budget: all revenues are simply
transferred to households.18 The budget collects revenue from the production sectors; there is no income tax in
the model. The only tax in the economy is a value-added tax levied at a rate of 100%. B is obligated to transfer

17
Note that in this example, manufacturing is actually not a value destroying sector. It loses money on
production because costs exceed revenues. But the value of output is equal to the value of purchased inputs, so it is
only a destroying value in the weak sense, see the appendix.
18
For even further simplicity of exposition, assume that the transfers are pensions.
FOUR SECTOR MODEL Page 6

all taxes collected to H but cannot borrow. Unmet budget expenditure obligations will be considered budget
arrears to H.

4. THE “VIRTUAL ECONOMY”, OR WHAT APPEARS TO HAPPEN WHEN EVERYONE

PRETENDS M’S OUTPUT IS WORTH 300

We now examine how this economy appears when virtual prices rather than actual market prices are
used to value output. Specifically, we assume that M prices its output at 300 rather than 100.19 G and B accept
this valuation. For now we treat the willingness of G and B to accept M’s overpricing as an assumption, but
clearly this is an essential point to be explained. To more clearly follow the transactions involved in this
economy, let us assume that the inputs (gas and labor) are supplied to M and production occurs before any
financial settlements are made.

An important consequence of virtual pricing is that it therefore appears, contrary to reality, that M is a
value-adder. As the output (cash value of 100; “virtually” priced at 300) sits at M, waiting to be shipped out,
the following claims prevail among the sectors:

H - Household Sector
Has For Amount Has For Amount
claims obligations
on sector to sector
M wages 100 None
B pensions 200
____________________________________________________________________

G - Gazprom
Has For Amount Has For Amount
claims obligations
on sector to sector
M gas 100 B taxes 100

19
The rate of overpricing of output in the virtual economy model is chosen arbitrarily here. The point is to
choose a price that makes M into an apparent value-adder. But, in fact, a rate of overpricing of 300% is roughly
consistent with what we know about many enterprises’ actual practice. According to the Karpov Report,
Commodity output used for the purpose of tax offsets is typically overpriced by a factor of two. Barter goods used in
transactions between enterprises are overpriced by a factor of 2-2.5. Promissory notes (veksels) may have a
nominal value up to five times the cash value. [Source: the Karpov Report]
FOUR SECTOR MODEL Page 7

M - Manufacturing Sector
Has For Amount Has For Amount
claims obligations
on sector to sector
H wages 100

None G gas 100

B taxes 100
______________________________________________________________________

B - Budget Sector
Has For Amount Has For Amount
claims obligations
on sector to sector
G taxes 100 H pensions 200
M taxes 100
______________________________________________________________________

These claims and obligations are the amounts each sector expects to pay and expects to receive. The
crucial point, however, is that there are insufficient resources to meet all these obligations. All that is available
in terms of value are goods (the output of M) worth 100 in cash if sold outside the system (exported or
domestic sales), but nominally priced at 300 for transactions between M, G, and B. Hence, we must now
examine how financial settlements are made in the virtual economy.

4.2 THE FINANCIAL SETTLEMENT STAGE IN THE VIRTUAL ECONOMY

To follow the settlements more clearly, we take each sector in turn.

We begin with sector M. Notice that because of virtual pricing M is no earning profits, so it has
acquired an obligation to the budget. It now has 300 worth of claims, equal to the 300 virtual value of its
output. Let us assume that the manufacturing sector allocates its output equally among all three sectors to
which it has debts. That is, it allocates output virtually priced at 100 rubles to each of G, B, and H. If the
virtual prices are accepted by G and B, this settles M’s debt of 100 to G for gas deliveries, and 100 to B for tax
payments. How about the wages of 100 which M owes to its workers? Although there are frequent cases of
workers being paid in kind in Russia today, we assume initially that the workers in this model must receive
wages in cash. Hence, M sells one-third of its output for cash to pay wages. Because the market value of M’s
output is 1/3 of its virtual value this means that workers receive cash wages of only 33. Thus M settles its
accounts with G and B, but it ends up with wage arrears to H of 67.

G’s settlement is straightforward. G takes the 100 (virtually priced) units of output it received from M
and remits it to B as its tax payment in full. Because B credits G’s tax obligations in full, at the virtual value, G
does not suffer from M’s over-pricing. Since G has no direct dealings with H, G’s accounts are balanced.
FOUR SECTOR MODEL Page 8

What about the budget sector? B has now received 100 (virtual) in goods from M and 100 (virtual) in
goods from G. Together, these have a cash value of 67, which B realizes and transfers to H. As a result, B’s
accounts with M and G are balanced. But B owes H 133.

Finally, households have received a total of 100: 33 from M, and 67 from B. H is still owed 67 by M
and owed 33 by B in budget transfers.

4.3 COMMENTS

When all is settled, H has received a total of 100, which was assumed to be its minimal survival
requirement. This means that from the households’ standpoint, the system is viable, and that it can thus start
again for another round. Of course, the minimum survival requirement in our model was arbitrarily set. In real
life it is not. It is the minimum amount of cash needed for H to survive. It determines the “cash constraint” for
the system.” There are interesting alternative ways to think of the cash constraint. We could, for instance, have
set the household survival requirement higher than 100. This would force the households to earn some cash
outside the system in order to survive. Or we could have kept the same 100 minimum for H and allowed outside
earnings to lower the cash requirement inside the system. Such an arrangement could be thought of as an
implicit contract between M and H to allow outside earnings on job time, for example.

In this version of the model, it is only the total of income received by H that matters. This implicitly
assumes that the household sector will redistribute amounts received in the form of wages and pensions to
ensure survival of all members of H. A complicating condition would be to allow B to choose to fulfill part of
its expenditure obligations in the form of in-kind transfers. It would then pass on some of M’s output in natural
form to H as budget spending (say, as hospital linens or subway construction...). This would complicate
matters, since H could not easily reallocate those transfers within H.

The virtual economy operates because of three key conditions. First, and most important, G pumps
value-added into the system. The rest of the economy just redistributes G’s 100, ultimately to H. Of course, in
the actual economy the lines between the value-adding and value-destroying sectors are more blurry.
Conceptually, however, what matters is that there is a value pump that is operating to support the rest of the
economy.

The second key condition for the operation of the virtual economy is that both G and B accept M’s
overpricing of its output at 300 rather than 100. This means that they underwrite M’s pretense that it is a
value-adding operation. Of course they will only accept this pretense if it is in their own interest to do so. We
have already seen that G will play along if it can use the inflated value of output to pay taxes.20

As for B the issue is more complex. Certainly budget authorities would in general prefer taxes paid in
cash rather than in goods with inflated prices. It is frequently the case, however, that governments are faced
with accepting taxes in kind or not collecting at all. Faced with this choice offsets are typically accepted; some
tax receipts are better than none. But this is not the whole story.

Offsets often arise out of a conflict between different agencies of government or between different
levels of government. That is, with not enough value to go around, offsets (payment of taxes in kind) may serve

20
This is clearly the case in our example where all G’s profits are taxed away. G is then indifferent as to the
form in which it pays taxes. In the actual case G’s willingness to accept virtually priced output depends on how
much taxes it owes. G will accept virtually priced output that it can pass on in tax payments.
FOUR SECTOR MODEL Page 9

the purpose of "tying" budget receipts to one particular recipient to the exclusion of others. Take, for instance,
the case of a conflict between ministries. When the Ministry of Finance restricts the budgets of important
ministries (e.g., defense, Minatom), rather than reduce purchases the ministries continue to order output but are
unable to pay. These debts are then canceled in lieu of the tax obligation of the enterprise. Tax offsets thus
become a means for one particular agency of the government to continue to acquire output when there is
insufficient revenue to support the expenditure of all.21

The third key condition is that there has to be a buyer outside the system who supplies cash by buying
some of M’s output—enough for M to meet its cash constraint—at the market price. That is, the system is not
self-contained.

4.4. THE REAL ECONOMY: NO ONE PRETENDS THAT M’S OUTPUT IS WORTH MORE THAN 100

This section compares the apparent outcome of the virtual economy to what is really happening
beneath the pretense. All that is necessary is to assume that no one, including M, pretends that M’s output is
worth anything other than the 100 it actually is worth. Note that we are not examining the equilibrium
adjustment of the system to the elimination of value destruction. Rather, we simply account for the actual value
of the flows in the system.

Again it is useful to begin with M. Because the value of its output is really only 100, in a non-virtual
economy M would have to report a loss of 100 instead of a profit of 100. It therefore would have no tax
obligation. With sales revenue of only 100, M could not pay both G (to whom it owes 100 for gas) and H (to
whom it owes wages of 100). It would have to apportion the 100 it does have between them.

The simplest assumption to make is that M pays equal shares to each of its creditors, that is, 50 to H
and 50 to G. This leaves M with wage arrears of 50 to H and inter-enterprise arrears of 50 to G. Of course,
these arrears numbers are arbitrary, depending on how M pays its bills. It is important to note that, however,
M’s payments are apportioned, the sum of payments will equal 100, as will the sum of the arrears.

Turning to the remaining sectors, G remits to B the 50 it receives from M. Since on a accrual basis G
owes B a total of 100 in VAT, this leaves G with tax arrears of 50. B’s only revenues are what it receives from
G, since M has produced no value added and thus owes no taxes. B transfers to H the 50 it received from G. B
still has budget arrears of 50. H survives because it receives 100 (50 from M and 50 from B).

It is important to recognize that these proportions could be changed, depending on how M initially
allots fulfillment of its obligations to G and H. Again, H always receives 100. Whatever M pays to G will
eventually be passed on, via B, to H.

21
Tax offsets are also accepted by local governments as a means of appropriating a larger share of taxes than
would be possible if taxes were paid in money. Taxes paid with money would be split between federal and local
levels according to sharing rules. If taxes are paid to the local government in kind this precludes transferring the
federal share to Moscow.
FOUR SECTOR MODEL Page 10

5. COMPARATIVE MACRO INDICATORS, VIRTUAL AND REAL:

It is interesting to compare how the economy appears to be performing in the virtual economy regime
with how it is actually performing according to the underlying reality, as we do in Table 1.22 This provides
some perspective as to why agents may prefer the virtual economy. Examining the results in Table 1, we obtain
a surprising result: On nearly all counts, the virtual economy’s aggregate performance indicators (sales, profits,
GDP, output) are going to look better than the real economy’s.

First, consider the budget. The planned budget in the real variant is only half the size that it is in the
virtual variant. What does that mean? The larger virtual budget is accompanied by larger arrears in relation to
planned budget outlays. Households are, in fact, no better off in the virtual economy. But moving to the real
economy means that (nominal) pensions are cut by 50%. Of course, in practice total household income is the
same in both variants, because value added actually produced is identical. In reality, nothing changes as we
move from the real to the virtual economy. But the pretense is what counts. In moving from the virtual to the
real variant, the perception will be: “You cut pensions in half!”

Turning to arrears, we have already noted that the arrears picture will depend on the decision by M of
how to allocate payments between H and G. It was assumed above that each received half of the cash value of
M’s product. In the general case, let the proportion used for H’s wage payments be “w.” Then G will receive
“100-w”. Final household income is always 100 (since that is still the total product that can be allocated). The
wage arrears will be offset by budget arrears, since what is paid to G (and passed through B to H) could always
be paid directly to H in wages. The greater the share of M’s product paid directly to H in wages, the greater the
amount of total arrears and the greater the number sectors that have arrears of one kind or the other. This is
evident from Table 2.

In the virtual variant, arrears problems exist between M and H (“wage arrears”) and between B and H
(“pension arrears”). In the worst real variant, there are now two qualitatively new kinds of arrears problems
between G and M (“inter-enterprise arrears”) and between G and B (“tax arrears”).

Finally, of course, the virtual economy masks the non-viability of M. In the virtual economy, M
appears to add value of 100. In the real variant, M is a clear loss-maker. Therefore, any attempt to shift from
the pretend virtual world to the honest real world would be very unpopular. It would mean slashing pensions,
irritating Gazprom by branding it as a tax delinquent and demanding more cash taxes from it, and threatening
the bankruptcy of the manufacturing enterprise and complete loss of jobs and wages for the population.

22
It is important to emphasize that this comparison is between the virtual economy and the underlying reality.
A further interesting comparison would focus on outcomes when full adjustment out of value destroying activity is
accomplished. Clearly in that case welfare would be enhanced. Our point, however, is to show why that
adjustment, which clearly represents a potential Pareto improvement does not occur.
FOUR SECTOR MODEL Page 11

TABLE 1: COMPARISON BETWEEN VIRTUAL AND

REAL VARIANTS

VIRTUAL REAL
Total sales 400 200
Total profits 200 0
Profit rate 50% 0%
Total value-added (GDP 300 100
Industrial output 400 200
Wages
Accrued 100 100
Actual 33 50
Budget
Revenues: Planned 200 100
Actual 200 50
Spending: Planned 200 100
Actual 67 50
Total household income
Accrued 300 200
Actual 100 100
Arrears
Wage 67 50
Inter-enterprise none 50
Budget 133 50
Tax none 50
TOTAL ARREARS 200 200
FOUR SECTOR MODEL Page 12

TABLE 2: VARIATION IN ARREARS DEPENDING ON

M’S ALLOCATION OF WAGE VERSUS GAS
PAYMENTS
(w = wage)

VIRTUAL REAL
General w=0 w=50 w=100
Total sales 400 200 200 200 200
Total profits 200 0 0 0 0
Profit rate 50% 0% 0% 0% 0%
Total value-added (GDP 300 100 100 100 100
Industrial output 400 200 200 200 200
Wages
Accrued 100 100 100 100 100
Actual 33 w 0 50 100
Budget
Revenues: Planned 200 100 100 100 100
Actual 67 100-w 100 50 0
Spending: Planned 200 100 100 100 100
Actual 67 100-w 100 50 0
Total household income
Accrued 300 200 200 200 200
Actual 100 100 100 100 100
Arrears
Wage 67 100-w 100 50 0
Inter-enterprise none w 0 50 100
Budget 133 w 0 50 100
Tax none w 0 50 100
TOTAL ARREARS 200 100+2w 100 200 300

6. WHY DO THE SECTORS PARTICIPATE IN THE SYSTEM?

The comparison between the virtual and real variants shows the incentives for the sectors to continue to
hide the underlying reality once they are in this system. But the question is, why be in it at all? We consider the
position of each sector in turn.
FOUR SECTOR MODEL Page 13

The motives of the M sector are the most obvious. If it were not in this system, it would likely be out of
business altogether. It is a value-subtracting enterprise. The director presumably gains utility from his position.
He will prefer to stay in an enterprise that everyone pretends is successful than to concede failure. It is the
survival motive which is the key motivating factor of enterprises in the M sector, and this is critical to the
operation of the virtual economy.23

The household sector is not being fully compensated for the labor it supplies to M. It might therefore
appear that it is supplying net value-added to the system (like G). Why does it then participate? The important
point to remember is that H is also receiving payments from B. In total, it is being compensated for the labor it
supplies to the system. If the system were to break down, H would be unemployed. The key question for
households concerns the alternative jobs that are available: Can members of H earn more than 100 elsewhere,
net of adjustment costs? Unless H believes it can be employed outside the system (in the “new” economy), it
will remain in the system.24

Gazprom’s reasons for remaining in the virtual economy are more complex. G is the net supplier of
value-added to the system. In our example, it would be no better off outside the system because we assume that
in any case it would have to pay VAT at a rate of 100% if it exported the 100 rubles worth of gas that it now
supplies to M. But this assumption is made only for convenience. The question thus remains, why does G stay
inside the system?

The fundamental reason why G continues to play is that it is able to keep some of its earnings. To use
the term we will discuss in more detail below, some of the value is permitted to leak out from the system into
the pockets of G’s owners. This leakage is sanctioned as a side payment to keep G pumping value into the
system. In the case of Gazprom it involves allowing it to keep export earnings. In a sense, sustaining the virtual
economy involves a sharing of the value produced by G. To better understand that, it is worthwhile considering
Gazprom’s place in the Russian economy.

6.1 THE PRIVATIZATION LOTTERY

As explained in Section 2, above, the Soviet economy was characterized by value creation in the
resource sector and value destruction in industry. The preferences of the leadership were biased towards
defense and industry, however, and it was important that it appear that these sectors were producing value. The
price system was adjusted accordingly. Because all property was state-owned the returns to specific assets were
irrelevant. Hence, arbitrary pricing resulted in artificial returns to activities in the Soviet economy.

23
See Ickes and Ryterman (1994) for a discussion of the theory of survival-oriented enterprises. A recent paper
by a pair of economists from the IMF and World Bank argues that “barter is also very much a survival response
engaged in by firms that do not have a ready market for their product. In this, they are aided by local governments,
which might prop them up through their procurement efforts...Increasingly the subsidy seems to be coming from
natural monopolies like Gazprom, RAO UES and local electric utilities, and the railroads” (Buckberg and Pinto
1997: 16).
24
And, of course, there are people who have left the virtual economy to find jobs in the new economy.
FOUR SECTOR MODEL Page 14

Because values were redistributed according to these preferences, and because the true nature of
production was distorted from view, agents had little idea of their true value in a market economy.25 But agents
derived their social identity, and sense of self-worth, from their position in the Soviet economy.

Privatization in Russia involved demarcating assets into specific bundles and assigning ownership to
them. In a market economy these assets would have different values than in the Soviet economy. This is the
essence of market reform, to improve the allocation of resources and make apparent the true value of resources.
But because the agents in the economy could have little idea of the market value of their assets, human or
physical, in the vastly different economy of market reform, privatization in Russia was tantamount to giving
agents a lottery ticket.

The outcome of the lottery, to simplify a bit, was that Gazprom was the winning ticket. This is not
surprising because we know that the resource sector was the value producing sector in the Soviet economy.
Now suppose that after the privatization lottery, property rights are completely enforced and that the
government absents itself from the issue of distribution. Then shareholders of Gazprom would have valuable
assets, and all other agents would hold assets with no market value. Call this the lottery distribution, L.

We can consider an alternative to the L-distribution. Suppose that Gazprom is broken up rather than
privatized, and that the assets are equally distributed to all citizens. Call this the E-distribution. This presents
the extreme opposite case to the L-distribution.

What happens when agents are unwilling to accept the L-distribution? There is no unique way to
understand how assets should be redistributed from the Soviet regime. Russia is a democracy, so we would
expect this to be a political issue. But if we assume that there exists some minimally acceptable level of
inequity among the Russian population, it follows that there will be limits on how much value owners of
Gazprom can obtain. As a result, there has to be a mechanism to redistribute some of the value produced by
Gazprom among the other lottery-ticket holders.

This means that we have yet another important way to conceptualize the meaning of the virtual
economy: It is a way of making Russia’s privatization lottery more equitable. It is a means of paying off on
more tickets. The redistribution of G’s value makes other tickets have positive, rather than zero, value. It is the
outcome of a bargaining game between agents in the economy over the value of the lottery tickets. The reason
that this game is played is because it is politically impossible to enforce the L-distribution in a democracy in
which equity norms differ substantially from those of the L-distribution. Consequently, Gazprom must pay off
other parts of society to retain its share of the value that is produced. This bargain determines where the
economy ends up between the L-distribution and the E-distribution. But because the L-distribution reflects the
market value of the assets, redistribution of value in the virtual economy represents the postponement of
reform.

25
This effect was enhanced by the predominance of the shadow economy in the Soviet Union. The cascade of
payments that spread through the system -- the “loot chain” as Gregory Grossman has called it in parallel with the
food chain -- meant that the income of a typical person was a convolution of formal and informal activity. This
makes it very hard to estimate how one would fare in a liberalized economy, because of the effect this may have on
the loot chain.
FOUR SECTOR MODEL Page 15

7. LEAKAGE

So far we have considered the virtual economy as a closed system. Value is pumped in by G to be
redistributed within the system, with part of that value being destroyed by M in the process. But there is
another way the system can lose value. Value may “leak” out.

Leakage from the virtual economy takes several forms. It can be legal or illegal, sanctioned or
unsanctioned. The value that leaks out may stay inside the country or may be transferred abroad (capital flight).
The most important distinction, however, is whether the leakage is good for the system or bad for it. Good
leakage can be thought of as a necessary cost of keeping some participants in the game. That would be one way
of thinking about the role of Gazprom. In our example, Gazprom contributes all of the value it produces to the
system. As a privatized company one would assume that its owners would prefer to export all the gas for hard
currency.26 But this is politically impossible. In practice, Gazprom is allowed to siphon off (and pocket) a
certain share as payment to keep it performing its role in the system.

While some leakage is thus good for the virtual economy because it makes the system work (like the
Gazprom cut), other leakage is damaging (like the theft of wage funds by an enterprise director, which makes it
more difficult to meet the cash constraint, or a manager’s diversion of cash from taxes). Viewed in the context
of leakage, the relationship between corruption and economic reform takes on greater complexity. Reducing
corruption is typically considered a key element in accelerating economic reform. In Russia’s virtual economy,
the opposite may in fact be the case. If reducing corruption results in less leakage from the system, more value
remains to support the continued operation of loss-making enterprises.

7.1 “A TAX CRACKDOWN”: THE GOVERNMENT REFUSES TO ACCEPT THE OVERPRICING OF

M’S OUTPUT

Anti-corruption is not the only intended reform goal that becomes complicated by the virtual economy.
Tax reform is even worse. Nonpayment of taxes may result from two causes. The first is the nature of the
virtual economy itself, which makes some nonpayment of taxes almost inevitable. The second is connected to
leakage.

In the pure virtual economy, there are no tax arrears. (Cf. Table 1.) But as soon as the government
demands that the virtual economy participants pay some taxes in cash, there will almost inevitably be tax
arrears even if, as the model assumes, no one actually evades taxes.27 To begin to understand the complexity of
tax reform in the virtual economy, we examine, in this section, the implications of a “tax crackdown” by the
government, i.e., a campaign to try and secure more cash to the budget. This turns out to be an intermediate

26
Notice that even if Gazprom faces constraints on its capacity to export in hard-currency markets, it still has
the option of keeping gas in the ground. Faced with the option of selling at a discount to domestic consumers or
keeping the resource for the future when domestic customers could pay a higher price or new export markets could
be found, a secure private owner would certainly choose the latter. This is especially so when tax liabilities
accumulate according to the virtual value of sales rather than the reduced value of discounted sales. As with the
case of greater exports, a policy of keeping gas in the ground would enhance the likelihood that Gazprom would be
re-nationalized.
27
It is only in the case that enterprise M uses all of its product to pay G and none to pay H, that there are no tax
arrears. See Table 2's “real”case where w = 0. But w = 0 means that M pays nothing to its own workers in wages.
FOUR SECTOR MODEL Page 16

variant between the virtual variant first described and the underlying real process. In the end it would likely be
perceived by the virtual economy’s participants as worse than both.

To see this, assume now the government decides it does not want to continue to accept tax
offsets/barter payments. M continues to price its output at 300 and G accepts that price. M therefore delivers
one-third of its output (virtual value 100) to G. As in the original scheme, M pays H the cash value of one-third
of its output, i.e. 33. M’s arrears to H are 67.

The key difference is that now B refuses to accept G’s and M’s offer to each remit 100 (virtual value)
worth of M’s output as tax payments. B demands that they remit their taxes in cash. M still has a tax obligation
of 100, since it continues to price its output at 300. It remits 33 in cash and has tax arrears of 67. G will also
remit 33 in cash and have tax arrears of 67. In the end, the tax crackdown variant will be the same as the real
variant except that in addition to all the aforementioned problems, M, too, will have a tax arrears problem. In
other words, this is worse in this respect than the real variant.

The situation will also appear to be worse than in the pure virtual economy. After the tax crackdown,
the wage arrears remain the same (67). The cash value of tax revenues to the budget is the same (67). The
budget arrears to H are the same (133). But now there is a major new problem which did not exist at all in the
virtual economy: tax arrears. Both producers in the system, the value-adder G and the value-subtractor M, are
declared to be tax delinquents. This despite the fact that the point was to improve tax discipline!

Of course, if the cash tax collection policy is pursued vigorously enough, the likely outcome is that the
budget actually will collect more value from M, but only because M pays less to either G or H, or both, in
order to have more cash to deliver to B. This means either increased wage arrears, or increased debts to G on
the part of M, or both. This is precisely what happened in the Russian economy in the first quarter of 1998.28
Once again, the outcome is completely predictable as long as one keeps in the mind that the cash tax collection
campaign did nothing to change the fundamental fact of the virtual economy: There is still not enough value
being created to meet all the claims. The tax crackdown just temporarily shifts the flow of value relatively away
from G and H and toward B.

The remarks above related to the tax crackdown’s effect on M. The impact on G may be quite
different. If G is earning additional cash on the side (that is, outside the virtual economy, as a form of ), the
government’s demand that G pay more cash will be tantamount to changing the rules for G’s participation in
the system: it will be the same as raising the net contribution that G makes. This could be accomplished, but
presumably only if the government had a credible threat of punishment or retaliation if G failed to go along (for
instance, dispossessing the current owners, whether through “nationalization” or “privatization”).

28
[Data on additional taxes collected versus increase in wage arrears.]
Representatives of Russian labor have pointed explciitly to the trade-off between tax arrears and wage
arrears. Mikhail Shmakov, the head of Russia’s Federation of Independent Trade Unions is calling for an article in
the Civil Code that would obligate enterprises to pay their wages before taxes. Vadim Borisov, the director of the
Institute for Comparative Labor Relations Research in Moscow, an organization that is helping the unions in their
campaign to eliminate wage arrears, said “the government’s recent efforts to squeeze more taxes out of industrial
enterprises to pay public sector workers was only worsening the problem of private sector wage arrears.....
‘Industrial enterprises are paying taxes before wages, they are moving more of their accounts into the shadow
economy, and they are increasing the use of barter. All these tendencies are continuing the demonetization of the
economy,’ he said.” [Financial Times, 9 April 1998, p. 2]
FOUR SECTOR MODEL Page 17

Of course, the idea with the crackdown may be that by forcing G and M to pay cash to the budget, this
will force G in turn to refuse to accept the inflated price of M’s output—i.e., to impose the full real variant. If
so, that would have even worse political consequences, including the possible bankruptcy of M. To avoid this
consequence, G only has to tolerate M’s arrears (possibly by borrowing money outside the system).29

8. THE ROAD TO RUIN

Our goal in this paper was to develop a simple model of the virtual economy. The essence of the virtual
economy stems from two factors: (1) An important share of the economy is not creating value, but destroying
it, and (2 ) most of the participants in the system pretend this is not happening. Value-subtraction and pretense
are core elements of this peculiar system. The result of these elements are the phenomena that draw most
attention: barter, veksels, and tax offsets. But these phenomena are merely mechanisms to make the virtual
economy operate.

It is especially important to underscore the role of pretense. Because of the pretense that there is more
value being produced than there actually is, there are exaggerated claims on the value that is produced. The
pretense is the root cause of the nonpayments difficulties that plague Russia. There is less value produced than
there are claims on it. In particular, this is the problem of Russia’s budgets. The apparent low rate of tax
collection on the revenue side and the failure of the government to meet its spending obligations on the outlays
side, most notably the unpaid wages and pensions, both result.

In our simple model the virtual economy can continue to reproduce indefinitely. The system appears to
be self-sustaining, as long as Gazprom continues to pump value into the system. Of course Gazprom is not the
only value producing sector in the economy. New private enterprise is also a value creating sector. But it is
critical to note that new private enterprise is not independent of the virtual economy. The continual need for
cash of the virtual economy is an inhibiting factor to the development of the new private sector. Taxation falls
heaviest on this sector which cannot pay taxes via tax offsets and other machinations. Because the virtual
economy needs an infusion of value to survive it acts as a check on the potential of the private sector to grow.
A functioning private sector is needed as a source of cash and value, but its potential is circumscribed by the
pressure that comes from the virtual economy.

29
As an example of how Gazprom can relieve pressure on the manufacturing sector by borrowing money
outside the system, consider this news: “Russian gas giant Gazprom is to press ahead with a bond issue this month worth
some one billion dollars as part of a borrowing programme needed to make up for customers' unpaid debts, ITAR-TASS
reported. ... Further unsecured and convertible bond issues will follow this year to increase the company's 1998 borrowing
to two billion dollars, [Gazprom chairman Rem] Vyakhirev said.”
“Last week, Deutsche Bank announced it would lead a syndicated loan to Gazprom worth 230 million dollars.”
AFP release May 5, 1998.
Comment: The government squeezes Gazprom to pay cash taxes so it (the government) doesn't have to borrow so
much on the bond markets. Also, so that Gazprom will squeeze its own customers to pay in cash, etc. But Gazprom turns
around and mortgages future gas exports by borrowing cash, thus refusing to tighten the cash constraint for the enterprises.
The result: no change at all! Gazprom borrows to keep the cash constraint eased instead of the government.
FOUR SECTOR MODEL Page 18

APPENDIX: A NOTE ON VALUE DESTRUCTION

A key element of the virtual economy is the continued existence of a value-destroying sector. The
distinction between value creation and value destruction is fraught with confusion. Hence it is useful to discuss
the issue.

The most basic notion of value destroying activity (negative value added) is that the market value of the
purchased resources used to produce output exceeds the market value of the output itself. Because value added
is the difference between the market value of output and the market value of purchased inputs, when this
difference is negative it means that the enterprise adds negative value to these resources. In such cases, the
enterprise has transformed the inputs in a manner that actually makes them worth less than what was started
with.

It is important to note that value destruction does not mean that the products are without value. An
enterprise may produce a useful product and destroy value; value destruction means that the purchased input
costs of producing the product exceed the market price. This is important to keep in mind, especially in the
wake of liberalizing external markets. Exposure to world prices may mean that domestic producers that were
able to cover their material costs at domestic prices may no longer be able to at world prices. Thus any
assessment of value destruction depends on the opportunities foregone.

Negative value added is a clearly a much more severe condition than lossmaking. An enterprise loses
money when the revenue (market value) it receives is less than the cost of production. Costs of production,
however, include non-purchased inputs (labor and capital services), so an enterprise can be lossmaking without
destroying value. Indeed this is normal: negative value added requires that losses exceed the cost of primary
inputs. Any enterprise that uses capital and labor services will incur costs in addition to the value of purchased
inputs. Negative value added is significant because it means that the application of labor and capital to the
inputs result in less value than what the process began with.

What happens when an enterprise is able to delay paying for its material inputs? Suppose that in any
period enterprise I pays only the fraction  < 1 of its material costs. Then it is possible that the enterprise is
financially solvent even if it is actually destroying value. This would be true when:
% i  wl
Pizi PMM < 0 < Pizi wl PMM (1)

where %i is profits of enterprise I, w represents the unit costs of primary inputs, l, Pi is the price of good I, zi is
the output level of the firm, and Pm is the price of materials, M. This is possible when material arrears are
larger than wage costs, wl < (1 ) P MM . Notice that in this case some other enterprise is essentially lending
to enterprise I. A natural case would be natural gas and the materials purchased are from Gazprom. This is the
canonical case considered in this paper. On average, of course, arrears on inputs and on sales will balance out.
But a truly profitable firm can subsidize the production of a value destroyer. Indeed, this is the essence of the
virtual economy.

Because SNVA is so strict it useful to work with a weaker notion of negative value added (WNVA). It
is useful to include the costs of replacing capital explicitly. Suppose that revenues are sufficient to cover
material costs and wages exactly. In this case the enterprise loses revenue at the rate of its capital costs.
Because revenues are insufficient to cover the depreciation of capital, the productive capacity of the enterprise
is decreasing over time. It may be useful to think of this situation—where revenues are insufficient to cover
depreciation—as one where value is being destroyed in the weak sense. The capacity to produce national
income, in this case, is transient. Over time national income will fall as capital must be replaced.
FOUR SECTOR MODEL Page 19

We must be careful with this notion. Consider the situation of an enterprise with large sunk costs
earning negative profits. Because the assets are sunk the opportunity cost of using these resources in production
is zero.30 As long as revenues exceed the variable costs of production it is advantageous for the firm to continue
production. In such a situation production adds to national income. But this is not a case of SNVA because
Vi > 0 in this case.

An interesting case in the Russian context occurs when the nominal wage rate exceeds the opportunity
cost of labor. This is important in cases where workers continue to work at enterprises even when wages are not
paid in full. Let w̃ be the true opportunity cost of labor, where w̃ < w . Then it is possible that at nominal
(virtual) wages the enterprise is a lossmaker but at the actual opportunity cost of labor it may not be:
1
Pizi wL PMM < 0 < Pizi w̃L PMM (2)

In this case the enterprise actually is profitable at the true opportunity cost of labor, although it is not at
nominal prices. Because the workers receive less than they are promised it is sometimes argued that these
arrears represent lending by the workers to the enterprise. In fact, however, if the enterprise pays the promised
level of wages (where they get the revenue is another story) this is actually a gift to them.
But this also means that at such an enterprise it may not be appropriate to speak about wage arrears.31 Labor is
not fully compensated because nominal wages exceed the true value of labor’s contribution. Note, however,
that even at the inflated level of wages the enterprise is producing value added.

It is important to emphasize that whether an enterprise produces value or not depends on market
opportunities. For example, an enterprise may produce value at domestic prices but destroy value at world
prices. If the economy is closed, production at this enterprise contributes to national income. Once the economy
is open, however, continued production destroys value. Many post-Soviet enterprises are in this position: the
change in relative prices caused by external liberalization renders enterprises that were value producing in the
Soviet period, value destroying in the post-Soviet period.32

Given the inherited production structure it could be the case that many (most) industries are not
producing value added at world prices. This may call for temporary protection, especially if the industries do
produce value at domestic prices. Normally, one might think that a sufficient devaluation of the domestic
currency would solve the problem of protected domestic sectors, because with a depreciated currency domestic
manufacturing will be more competitive because of the increased cost of imports. But if technologies are rigid
and energy dependent,33 then both material prices and manufacturing prices will rise in tandem; relative prices
and hence (negative) profitability will be unaffected. This problem is exacerbated when a large share of the

30
Assuming that there are not operating costs of using the capital.
31
But if we think about the promised excess portion of the wage—the part over and above the true value of
labor—as being part of the redistributionist “deal” that the virtual economy represents, then it becomes more
complicated. There may be some extra amount to the true wage that ought to be regarded as “equitable.” Arrears
are tolerated to a certain level, but not beyond. This is analogous to the deal with Gazprom, how much of its export
earnings it is allowed to keep, how much it subsidizes the rest of the economy, etc.

32
This is especially so in the defense sector. In the Soviet period the output of this sector was highly valued by
end-users (the Party), so the value of production exceeded its costs. With liberalization, however, the value placed
on this output has fallen dramatically.
33
As in McKinnon's substitution model (1993).
FOUR SECTOR MODEL Page 20

economy is in heavy industry and manufacturing. This may suggest a role for temporary protection until the
viability of different sectors can be assessed.34

It is also important to note that the response to value destruction must depend on how transient is the
phenomenon. When prices are liberalized it may be the case that value added is negative at some enterprises
because production must adjust to the new world prices. What is critical to note about Russia, however, is that
negative value added in 1998 can no longer be ascribed to transient lack of adjustment. Industrial production
has, according to official statistics, shrunk by 50% since 1991. Some of this decline in production must be the
abandonment of production that was destroying value. But in 1998 any value destroyers that still exist cannot
be due to slow adjustment; rather, it is due to the active adaptation of enterprises to the virtual economy.

Russian enterprises suffered two types of shocks that affect the calculation of value added. Internal
liberalization—freedom for agents to choose what to produce and consume independent of central planners,
thus setting domestic prices freely, etc.,—and (2) external liberalization—introducing world market prices.
Because these changes occurred at the same time we do not really know what “Russian” (free market) domestic
prices were/are, so we don’t know which enterprises might have been value-adding at these never-existing
domestic prices. It could thus be argued that by combining the internal and external liberalizations, reform
created too big a shock to enterprises. It may have made the distance needed to catch up to the world market too
large for most enterprises. Realizing that, enterprises may have given up on adjustment, and retreated inward,
eventually ending up in what we now call the virtual economy.

Suppose, instead, that there had been only internal liberalization, while fairly heavy protection against
the world economy was maintained. Might not then more enterprises have attempted a first stage of adjustment,
thus at least giving us a better sense for deciding which would be viable in a second phase of adjustment, with
or without protection?

It is important to note that negative value added occurs fundamentally because of a lack of property
rights. Some agents are unable to enforce the their rights to income streams. It is this conflict which keeps value
destroyers operating.

34
Hughes and Hare (1992) estimate the share of output in industries with negative value added (at quality
adjusted world prices) at Bulgaria (50.8\%), Czech Republic (34.8\%), Hungary (34.6\%), Poland (8.4\%), and
Soviet Union (22.3\%).
FOUR SECTOR MODEL Page 21

REFERENCES

Bosworth, B., and G. Ofer, Reforming Planned Economies in an Integrating World Economy. Brookings,
Washington, DC, 1995.

Buckberg, E., and B. Pinto, “How Russia is Becoming a Market Economy: A Policy Maker’s Checklist,”
unpublished mimeo, September 1997.

Gaddy, C. G. and B. W. Ickes., “Russia’s ‘Virtual Economy’.” Foreign Affairs, Vol. 77, number 5,
September/October, 1998.

________, and ________, “To Restructure or Not to Restructure: Informal Activities and Enterprise Behavior
in Transition,” Davidson Institute Working Paper, May 1998.

Ericson, R. E. “Priority, Duality, and Penetration in the Soviet Command Economy,” RAND Note N-2643-NA
(Santa Monica, Calif.: RAND Corporation, December 1988).

K. Hendley, B. W. Ickes, and Randi Ryterman, “Remonetizing the Russian Economy,” in H. G. Broadman,
ed., Russian Enterprise Reform: Policies to Further the Transition, The World Bank, Washington,
DC, November 1998.

Ickes, B. W., and R. Ryterman, "From Enterprise to Firm: Notes for a Theory of the Enterprise in Transition,"
in Robert Campbell and Andrzej Brzeski eds., Issues in the Transformation of Centrally Planned
Economies: Essays in Honor of Gregory Grossman, Westview Press, 1994.

McKinnon, R.I., The Order of Economic Liberalization. 2nd edition, Baltimore, Johns Hopkins University Press, 1993.

Thompson, W., “The Price of Everything and the Value of Nothing? Unraveling the Workings of Russia’s
Virtual Economy,” mimeo, Birkbeck College, September 1998.