Revised version of the presentation at the conference Post-Communist Russia in the Context of World Social and Economic Development at the Institute of the Economy in Transition in Moscow, December 1-2, 2000.
This paper assesses the development of real output during the initial transition in East-Central Europe and the former Soviet Union. It compensates for four factors, contraction prior to marketization, increased underreporting of output, the reduction of value detraction, and the elimination of implicit trade subsidies. Everywhere the decline in output has been much smaller than perceived. The Soviet economy was in a far worse shape than generally understood. The differences between failures and successes remain vast. The correlation between economic performance and structural reform becomes much stronger. This statistical disinformation was a major cause of flawed economic policies.
Toward the end of communism, output plunged in virtually all Soviet-type countries, according to official statistics. Turning dramatic in the first year of transition, it continued for years.1 Poland was the first to return to growth after two years of transition, while Ukraine did so only after eight years. The total registered declines in GDP range from 13 percent from 1989 to 1992 in the Czech Republic to 77 percent from 1989 to 1994 in Georgia. This has been widely proclaimed the worst depression in the industrialized world, exceeding the Great Depression of 1929-33.
However, both communist and post-communist statistics are deeply flawed, but in different ways. While everybody recognizes these statistical problems and some authors detail them, all proceed to work with official statistics, as no full alternative set exists. For many purposes, this approach is reasonable, but the fundamental question about the fate of real output is left unanswered.
The purpose of this paper is to figure out what really happened to output during the initial transition in the former Soviet bloc, comprising the former Soviet Union (FSU) and six East-Central European countries. We focus on four aspects, namely contraction prior to marketization, increased underreporting of output, the reduction of value detraction, and the elimination of implicit trade subsidies. We shall touch upon defense production, investment, and economic welfare, but the issue of this paper is real output in the period of 1989-95 for East-Central Europe and 1991-95 for the FSU.
Our conclusions contrast sharply to the conventional view. First, everywhere the decline in output has been much smaller than perceived, and a few countries experienced instant growth rather than contraction. Second, the Soviet economy was in a far worse shape than generally understood. Third, even after revision, the differences between failures and successes remain vast. Fourth, the correlation between economic performance and structural reform becomes much stronger after statistical revision. Fifth, flawed statistics have disinformed policy makers in post-communist transformation, inciting them to adopt inefficient gradual reforms, which reinforced rent seeking and prolonged stagnation. Economic welfare has diminished far less than output. Yet, no precise knowledge of the actual development of output in transition is possible because of paramount methodological problems.
The Focus Should be the Post-Communist Fall in Output
With the collapse of communism, officially-recorded output plummeted throughout the post-communist world. Annual falls over 10 percent were standard, and in Armenia GDP sunk the most with 53 percent in one year (see table 1).
[Table 1 approximately here]
However, statistical biases are monumental. The first problem is the starting point. Economic chaos prevailed at the end of communism, and Romania and the Soviet Union registered sharp falls of output in the last year of communism, 7.9 percent in 1989 and 6.1 percent in 1991, respectively (see table 2). While East-Central European transition is measured against the last communist year, the standard for the former Soviet republics (FSRs) is 1989, although it should be 1991, if we discuss post-communism. That correction eliminates an average of 5 percent of 1989 GDP of the decline for the FSRs.
[Table 2 approximately here]
Then, the registered contraction was 17 percent of GDP in Central Europe from 1989 to 1992, some 30 percent in Bulgaria and Romania from 1989 to 1997, and in the FSU an average of 40 percent, ranging from 18 percent in Uzbekistan to 65 percent in Georgia (see table 3). Five countries (Armenia, Azerbaijan, Georgia, Tajikistan, and, to a minor extent, Moldova) were hurt by military conflicts, but for most other countries these recorded drops were unparalleled in peacetime.
[Table 3 approximately here]
Sharp Increase in Unregistered Output
Central planning was a system of cheating. Everybody had an interest in over-reporting production, as bonuses of ministers, managers, and workers depended on their gross production. This led to persistent over-reporting, probably amounting to some 5 percent of GDP (Åslund 1990). The interest in such doctoring of numbers disappeared immediately with transition.
Under capitalism, on the contrary, people and enterprises are anxious to avoid taxes, implying a downward bias. Furthermore, statistical agencies failed to keep up with myriad new enterprises. Even in Hungary, enterprises of less than 50 employees were not included in aggregate statistics for years. A large unofficial economy emerged, which was not necessarily illegal, but just not reported to the state statistical office (Johnson, Kaufmann, and Shleifer 1997, p. 173).
Admittedly, an underground economy existed also in the Soviet Union, but it was tiny because of severe repression, as evident from the pernicious shortages. On the basis of interviews with Soviet émigrés in the early 1970s, Gur Ofer and Aaron Vinokur (1992, p. 100) concluded that private activity in the urban consumer sector would add just 3-4 percent to the Soviet GNP.
The only comparable GDP numbers available for many transition countries are based on electricity consumption, assumed to develop broadly in line with GDP (Johnson, Kaufmann, and Shleifer 1997). Table 4 shows the most elaborate and comprehensive estimates of the unofficial economy ranging from 27 percent of GDP in Hungary to 6 percent in Czechoslovakia and 12 percent in the Soviet Union in 1989 (Kaufmann and Kaliberda 1996).2 This method only approximates the development of the unofficial economy, and it cannot be applied to four countries in the region, and the series ends in 1995.3
[Table 4 approximately here]
With the start of transition, the underground economy expanded everywhere. Soon, however, it shrank both in successful reform countries and the most repressive state-controlled economies, while continuing to grow in partially-reformed economies. Hence, the unofficial economy peaked in 1991 in the most successful transition economies (Poland, Hungary, and Estonia), while in less reformist countries (Russia, Ukraine, and Azerbaijan) it was still rising in 1995. Mostly, the unregistered economy peaked when the official GDP hit its nadir.
On the whole, the unofficial economy expanded tremendously. The average unregistered share of real GDP in former Soviet countries rose from 12 percent in 1989 to 36 percent in 1994. In the extreme cases of Azerbaijan and Georgia, it exceeded 60 percent of total GDP, and presumably also in war-torn Armenia. In East-Central Europe, by contrast, the unofficial share rose from 17 percent in 1989 to 21 percent in 1992 but then dwindled to 19 percent in 1995.
Taking the unofficial economy into account, the economic development of the region looks very different (see the last two columns in Table 4). First, on average the contraction from 1989 to 1995 was 32 percent rather than 40 percent for the whole region, and 36 percent instead of 54 percent in eight CIS countries (compare Table 3). Second, the differences between the most successful reformers and the laggards are reduced substantially, as the unofficial economy grew most in intermediate reformers, such as Russia and Ukraine. This adjustment eliminates 18 percent of 1989 GDP of the purported decline in output in the CIS, and it is huge for some countries: for Azerbaijan - 39 percent of 1989 GDP, for Ukraine - 28 percent, and for Russia - 25 percent. Third, the underground economies in the most repressive economies (Belarus and Uzbekistan) shrank. With this single adjustment, the intermediate reformers Russia and Ukraine both overtake non-reforming Belarus, and Russia almost catches up with Uzbekistan, which seems eminently plausible.
Revisions of official GDP are undertaken all the time, considering not only output but also the end-use side of GDP (consumption, investment and net exports; Koen 1995). Gradually, they include ever more of the hitherto unregistered economy, and almost all revisions boost output numbers. Some of these adjustments have been incorporated in later statistical revisions. For instance, the first official report stated that Bulgaria's GDP had fallen by 26 percent in 1991 and by 22 percent in 1992 (ECE 1993, p. 73). Both numbers were later revised ? to half (12 percent) for 1991 and one-third (7 percent) for 1992 (see table 1). In 1999, Lithuania revised its national accounts radically, reducing its total decline from 1989 to 1993 from 62.8 percent (ECE 1998, p. 65), to 39.8 percent (ECE 1999, p. 199) and thus eliminating a purported drop of 23 percentage units!
The statistics of the five war-torn states (Armenia, Azerbaijan, Georgia, Moldova, and Tajikistan) are especially poor, as their statistical systems simply collapsed, together with registered output. Much of the strong recovery in Georgia in 1996 and 1997 appears to be rooted in the registration of previously-unrecorded economic activity. Turkmenistan statistics are a joke and best overlooked.4
Elimination of Unsalable Output or Value Detraction
The fundamental problem with socialist economies was qualitative. Enterprises had little or no interest in producing what customers wanted because of prevailing shortages of goods and services, as well as soft budget constraints on enterprises. The persistent shortages implied extreme monopoly, reinforced by severe protectionism. Enterprises aimed at attaining their physical production targets, happily ignoring quality and choice of products, which steadily grew worse. Almost anything was difficult to buy in the Soviet Union, and a typical Soviet grocery store was empty when communism collapsed. Richard Ericson (1994, p. 195) has perceptively characterized this state of affairs: "Thus the whole economic system was based on economic illusion?the pursuit of goals unrelated to economic value creation in the absence of real economic information." Partial market economic reforms had improved the situation significantly in Central Europe, notably in Poland and Hungary, but it remained bad.
Much of Soviet manufacturing was sheer value detraction, as Ronald McKinnon (1991) put it. For instance, Soviet fishermen caught excellent fresh fish. Rather selling it on the market, they processed it into often inedible fish conserves, reducing the fish's value to almost zero. Incorrectly, this value detraction was recorded as value added in national accounts and thus included in the GDP. Value detraction increased down the processing chain. Soviet raw materials were excellent, Soviet intermediary goods (such as metals and chemicals) were shoddy, while consumer goods and processed foods were substandard. Value detraction also involved excessive costs because of obsolete equipment still in use and uneconomical location, with heavy industry located far from both inputs and markets, producing what nobody wanted to buy in any case (McKinsey Global Institute 1999). Many unsalable goods disappeared in storage or were quietly scrapped without any statistical recording.
Proper national accounts should exclude most of the "production" of consumer goods and processed foods, and any elimination of such value destruction is positive. The decline in manufacturing was staggering everywhere, for instance, in Russia from 1991 to 1996, 84 percent in light industry, 44 percent in food processing, and 57 percent in civilian machine-building (Goskomstat 1997, p. 336). As it was difficult to find any manufacture goods that were really worthwhile to buy even at extremely low prices, this decline in manufacturing output seems to reflect some reduction of value destruction. Yet, it was recorded as a decrease of GDP, and most observers misconceived it as a major tragedy. The positive effect can be noticed in expanded exports of raw materials and intermediary goods, which have typically led economic recovery in transition countries. This is probably the greatest statistical confusion in post-communist transition.
Value detraction can be assessed in various ways. Unfortunately, we cannot calculate the eliminated value detraction directly because manufacturing?s share of GDP is not available. Another measure is trade with non-market economies as a share of GDP, but all socialist trade was not useless, and it cannot be easily related to GDP because of sharp swings in real exchange rates and thus GDP in dollar terms. The same is true of increased exports of raw materials and intermediary goods. One single measure is preferable to avoid double counting; it should be related to GDP in domestic currency; and it must be widely available. Rather than total value detraction, we are interested in eliminated value detraction, as much has been maintained for long through regulations and state subsidies.
The most relevant overall measurement of reduced value detraction available appears to be reduced over-industrialization, measured as the decline in the industrial sector?s share of GDP (see Table 5). It is reasonably neutral to GDP level and exchange rates, while reflecting a major structural improvement. Yet, this is a partial measurement. Although most value detraction pertained to manufacturing, it existed throughout the economy. Value detraction persists in non-reforming countries, while new production has arisen in parallel, but we need a long period of measurement to capture the whole adjustment.
[Table 5 approximately here]
For most countries, this decline in industrial share ? or reduced value detraction in industry ? is in the range 9-20 percent of GDP till 1995.5 This decline largely corresponds to the intensity of structural reforms. As hard budget constraints started to bite later on in most FSRs, the contraction of their industrial sectors continued after 1995, while non-reforming Belarus pumped up its old industrial sector after 1995, undoing its initial reduction of value detraction. It appears plausible that the share of unsalable goods, or value detraction, amounted to around 20 percent of GDP in the last year of communism in most countries.
Reduction of Implicit Trade Subsidies
The economic distortions of communism were especially severe in trade among socialist states, as such trade was largely politically determined, with regard to both commodity structure and prices. Socialist states mostly exchanged goods nobody wanted, forcing substandard and overpriced merchandise upon one another. The wrong things were traded for the wrong reasons between the wrong people in the wrong places at the wrong prices.
The share of unsalable goods in their mutual trade was probably even greater than in the domestic economies. For instance, Hungarian losses of exports to formerly-socialist countries consisted predominantly of machinery and buses, which Hungary hardly exported to the West (Gács 1995, pp. 165-6). Some enterprises had three lines of production: a high-quality line for free international markets, an intermediary line for the domestic market, and a substandard line for socialist partners. Much of the intra-regional trade consisted of exports of manufactured goods from the more developed countries to the energy exporters, which effectively paid subsidies to the exporters of manufactures.
Raw materials, on the contrary, were fine, but their low prices involved huge implicit export subsidies paid by the energy exporters, essentially Russia, Turkmenistan, Kazakhstan and Azerbaijan. The early decline of intra-regional trade amounted to an elimination of implicit trade subsidies rather than a costly deterioration of terms of trade, as the early literature on the collapse of the socialist trading system argued.
Berg et al. (1999) note that high trade dependence had the greatest adverse aggregate effect on the initial output decline. EBRD (1999) and Popov (2000) rightly specify the problem as trade with other communist countries, which was even more distorted than domestic trade. The decline in mutual trade between the post-communist countries was largely a beneficial shake-out of unsalable goods or unaffordable waste of raw materials, although a certain disruption of viable trade occurred. Trade restructuring comprised a desirable systemic change and the elimination of implicit trade subsidies. While the losses of implicit subsidies were real, they were inevitable costs of national independence. To avoid double counting, we measure the reduction of unsalable output only through the diminution of the industrial sector, while we consider implicit subsidies in intra-regional trade.
In 1991, the clean dissolution of the CMEA (Council of Mutual Economic Assistance) eliminated both unsalable goods and energy subsidies. Economists calculated the "costs" or changes in terms of trade for South-East and Central Europe, which pursued about half of their foreign trade with CMEA countries (Rodrik 1992; Rosati 1995; Gács 1995). Their assessments of the impact of the Soviet trade shock ranged from a high of 7.8 percent of GDP for Hungary (Rodrik 1992) to 1.5 percent for Czechoslovakia and negligent for Romania in 1991 (Rosati 1995, p. 152; see table 6). These totals are likely to be understated, as trade with market economies was enormously dynamic, providing a strong positive effect.6 The trade effect was greater on countries that traded more with the Soviet Union and the CMEA (notably Bulgaria), countries that were more open (most of all Hungary), and countries that imported a lot of energy (Bulgaria and Hungary). Thanks to far-reaching early liberalization of foreign trade, the East and Central European countries, including Estonia and Latvia, achieved shares of exports to the EU predicted by the gravity model as early as 1994 (EBRD 1999, p. 91).
[Table 6 approximately here]
Foreign trade distortions were far greater in the Soviet Union than in Central Europe. Extreme protectionism forced the Soviet republics to pursue 90 percent of their trade with one another. Further aggravating the situation, the CIS countries undertook slow trade and payments reforms, maintaining much of their mutual trade in unsalable goods till 1994. The share of mutual trade among the CIS countries dwindled gradually, from 57 percent of their total trade in 1992 to 33 percent in 1997 (Michalopoulos and Tarr 1997), more than the gravity model would have predicted (EBRD 1999, p. 91).
Lucjan Orlowski (1993) and David Tarr (1994) have calculated implicit trade subsidies for the FSRs, comparing the prior prices with prevailing world market prices. Orlowski dealt only with interrepublican subsidies, while Tarr included also subsidies in trade with other former socialist countries. Both focused on 1990 and their numbers are surprisingly similar (see Table 7). For seven FSRs the total effect was less than five percent of their GDP. Three countries exporting oil and natural gas provided substantial subsidies as a share of their GDP, namely Russia (17.7 percent of GDP), Turkmenistan (19.5 percent) and Kazakhstan (7.4 percent). These three countries benefited greatly from the abolition of implicit trade subsidies. Five states enjoyed substantial trade subsidies, namely Moldova (16.1 percent of GDP), Estonia (12.7 percent), Latvia (11.3 percent), Lithuania (9.7 percent), and Armenia (7.6 percent). Not surprisingly, these countries with the exception of Estonia have suffered comparatively large falls in output, although most have undertaken substantial reforms.
[Table 7 approximately here]
Subsidization dwindled only gradually after the break-up of the Soviet Union at great Russian expense. The IMF (1994, p. 25) has estimated the costs of direct Russian financing of the other CIS countries at 9.3 percent of Russia?s GDP in 1992 and the implicit trade subsidy at 13.2 percent of GDP. Thus, Russia's total burden was an unaffordable 22.5 percent of GDP or $18 billion in 1992. In absolute dollars, however, Russian financing plunged. Formally, the gains of other CIS states were enormous. Only direct credits ranged from 11 percent of GDP in Belarus and Moldova in 1992 to 91 percent of GDP in Tajikistan (see Table 8). The Russian government gradually reduced both its financing and implicit trade subsidies by raising commodity prices. Therefore, we want to avoid making an assessment for the immediate post-communist years, but by 1995 these subsidies were small.
[Table 8 approximately here]
In Soviet times, direct budget transfers between states were of limited significance, but they were substantial for Soviet Central Asia, whose five states benefited from large direct budget transfers from the central Soviet government. Lucjan Orlowski (1995, p. 66) has dug out these numbers for 1989, when Kyrgyzstan received 7.8 percent of its GDP in union budget transfers, Tajikistan 8.2 percent, Turkmenistan 9.0 percent, Kazakhstan 9.3 percent, and Uzbekistan 11.3 percent of its GDP. From 1994, however, these subsidies were gone. These were inevitable losses for the Central Asian republics, connected with their independence rather than any change of economic system. Presumably, these subsidies were not included in their official GDP, so we make no compensation for them, but obviously their elimination hurt economic welfare in Central Asia, especially the provision of public services. The previous donors, primarily Russia, benefited when these transfers ceased, but these subsidies were presumably included in the donors' GDP.
Thus, the foreign trade "shocks" reflect a combination of unsalable goods, previously-disregarded transportation costs and the elimination of implicit trade subsidies ? essentially from Russia, Turkmenistan and Kazakhstan ? to other countries. As these subsidies were implicit, they boosted the GDP of the receiving countries. Their elimination was a result of political independence rather than any cost of transition. Therefore, the implicit subsidies should be deducted from the base GDP of the former recipients to facilitate a comparison with their post-communist output, while they were presumably included in the donors' GDP, warranting no adjustment of their GDP. Because of the very gradual transition in the CIS, we avoid the years 1992-94 (Olcott et al 1999).
Collapse of Defense Production and Consumption
Soviet defense expenditure was a persistent dispute in the Western Sovietological community. Gradually, the CIA raised its assessment of Soviet defense spending to 15-17 percent of GDP in 1986 (Berkowitz et al. 1993), but that was based on the CIA?s clearly exaggerated estimate of Soviet GDP. As late as 1990, the CIA considered Soviet GDP per capita no less than 43 percent of the US level in purchasing power parities (PPP). The European Comparison Program (ECP), which cooperated with Soviet statistical authorities, undertook a careful empirical analysis, setting Soviet GDP per capita at 32 percent of that of the USA in 1990 (and Soviet household consumption per capita at only 24 percent of the US level; Bergson 1997).7 If we use the CIA assessment of Soviet defense expenditures and the ECP assessment of Soviet GDP, the defense burden would amount to 22 percent of GDP.
Yet, even these GDP numbers are too high, as the poor quality of goods and services cannot be fully considered, while shortages and forced substitution are disregarded. Thus, the Soviet Union probably spent about one quarter of its GDP on military purposes in the late 1980s (Åslund 1990), going to both military production and military consumption, but representing a sheer waste of public resources.
The Russian reform government swiftly reduced military spending to an internationally normal level of about 3 percent of GDP, while most other post-communist countries reduced such expenses to 1-2 percent of GDP. This reduction of defense expenditures resulted in a nominal decline in the 1989 GDP of about 22 percent in the whole FSU. Yet, this might be an exaggeration. Much of barter, arrears and enterprise subsidies pertains to the military-industrial sector. Western intelligence argues that a couple of percent of GDP should be added, because the military do not pay for all the cost they actually cause society, such as electricity and land usage, a counterargument is that the military may use more resources for black market activities than for defense.
Unfortunately, we do not possess sufficient information to distribute the military costs among the FSRs. For Russia, Belarus and Ukraine, this nominal decline must have been disproportionately large, because they had hosted most of the military-industrial complex, probably in the order of 20 percent of GDP, while about 10 percent of GDP seems reasonable for the other FSRs. In East-Central Europe, military expenditures were not much higher than in the West, but even there the trimming of the military sector probably accounted for a couple percent of the fall in recorded GDP. Yet, we abstain from making any correction here to avoid the accusation of double counting, as part of the declining defense costs are reflected in the contraction of industry. Still, an additional deduction of 10 percent of GDP for Russia, Belarus and Ukraine would seem justified.
In this paper, we focus on production, but the utilization of output needs to be kept in mind. Socialism was a system of waste. Soviet production usually needed three times more inputs than a Western factory, since costs were irrelevant to managers. Some of these losses represented inefficiency, others theft. With the introduction of harder budget constraints, enterprises started bothering about costs, sharply reducing domestic demand for inputs, such as steel, metals and chemicals. Initially, however, budget constraints were soft or lacked credibility, prompting energy intensity to rise everywhere.
The same was true of investment. Communist regimes prided themselves on huge investment ratios, but the socialist landscape was scarred by unfinished construction projects (Winiecki 1988, 1991). One reason was the accepted practice of theft by state employees from construction projects to build their own houses or repair their apartments. Enterprises also used unfinished construction projects to pressure the government to provide additional state funds, as the state usually financed investment. Therefore, the persistently high investment ratios in fixed investment were indications of theft and waste rather than substantial real investment. As ample capital goods were under-utilized or unusable, a contraction of investment for a few years was desirable to stop the notorious theft by employees, to halt the hoarding of investment goods, and to allow for a re-allocation of unused capital goods.
Socialist countries piled up large inventories predominantly of inputs, such as raw materials, which were labeled investment in national accounts. As these inventories continuously accumulated without any cyclical tendency, this was obvious waste. Poland had the best statistics, showing that "investment" in inventories amounted to 7 percent of GDP or one quarter of total investment in the mid-1980s.
Already in the early transition, reformers managed to introduce a demand barrier in a few countries, notably Poland, Czechoslovakia, Estonia and Latvia. The national demand curve shifted permanently, initially reducing recorded output. Substantial dishoarding of inputs and capital goods started, as desired, while stocks of finished goods rose to a lesser extent, reflecting the problems to sell leading to the characteristic over-production of capitalism. The dishoarding of inputs led to a stark decline in demand for enterprises producing inputs. Andrew Berg (1994) has calculated that the total reduction in inventory accounted for two-thirds of the total decline in Poland?s GDP in 1990. Yet, although Polish enterprises faced a real demand barrier in 1990, heavy manufacturing and mining contracted the least, suggesting that the budget constraints of large Polish producers remained pretty soft. Apparently, even Poland needed a more severe monetary crunch.
The investment that was sheer waste should preferably be deducted from GDP. A comparison with East Germany is apt. The German Institute of Economic Research (e.g. DIW 1977) in West Berlin assessed that East German GDP per capita was stably about 60 percent of the West German level, and the GDR had a higher investment ratio than West Germany. When the wall fell, it became obvious that the East German fixed capital per capita was only 30 percent of the West German level (Siebert 1992, p. 39).