The usefulness of even short term forecasts appears to be doubtful. The annual rate of decline of 1994 Russian GDP at the beginning of 1994 was predicted by most experts to be 6-8 percent; the actual rate is estimated to be twice as large. The 1994 ann ual CPI index for Russia was predicted to be more than 500 (EIU, Feb. 28, 1994); its actual level was 336%. At the beginning of 1994 the Yugoslav GDP was expected to decline by about 10% (Economic Commission of Europe, Spring 1994); according to EIU (4t h quarter 1994) it actually increased by a few percentage points.
Poor compatibility of Eastern European economic data with the Western
statistics.
For example, many official national accounting statistics in the former
Soviet republics do not include so-called nonproductive services; the
Czech Republic does not follow international standards in its
unemployment statistics; etc.
Problems pertinent to all economies that undergo fundamental structural transformations.
Compilation of internally consistent time series data for these economies is extremely difficult, if at all feasible. Simultaneous radical shifts in relative prices,
economic structure, technology, income and consumption patterns, quality and availability of consumer goods, and institutional changes make intertemporal analyses very difficult*
.
Incomplete information, its low quality and reliability. A well known problem, faced by all East European countries, is the existence of the unofficial sector (the "second economy") which is estimated at 10-30% of total GDP of these countries and which i s dynamically expanding. As a rule, the official statistics do not account for the activities of this sector, despite its substantial effect on the economy. Official statistics suffer also from a number of other omissions and distortions. They are ofte n published with significant delays or are not published at all. Discrepancies among the figures provided by different sources are very significant. The abrupt changes in economic indicators are often purely "statistical" phenomena. For example, when t he official exchange rates are replaced by market exchange rates, the economic status of a country changes dramatically and this change may have little to do with real output. (See the section below on output measurement.)
Measurement problems caused by price distortions and inflation. During the transition period, old problems of Soviet-type economies (tightly regulated markets, distorted prices, problems with exchange rates, the so-called soft budget constraint policies, etc.) are combined with new transition-related problems (deficit, indebtedness, and a high inflation). An accurate calculation of any economic indicator in an economy with four digit inflation becomes an impossible task.
According to many economists (Jan Vanous of Washington DC's PlanEcon, Jeffrey Sachs of Harvard, and others), official Eastern European statistics tend to overestimate the economic decline in the early 1990s and underestimate the speed of the current recov ery/growth. Recently PlanEcon began publishing its own statistics that include upward corrections to the official economic figures to take into account the activities of the second economy and to adjust for an "upward bias in the official GDP deflators". In light of these corrections, the economies of East-Central Europe look much brighter. Not only the current levels of output are higher than those officially estimated, but also the current growth rates are very high--according to PlanEcon they are c omparable with those of the East Asian "tigers". The following table provides examples for these PlanEcon corrections:
1994 GDP, 1989=100
Official PlanEcon
Czech. Rep. 81.4 93.5
Poland 91.9 115.5
Slovakia 82.4 94.7
Privatization.
No accurate data on the ownership changes is available. We have to
content ourselves with rough estimates that at best can be used as
proxies for a "true" privatization which by its very nature is a complex
process. Often it is difficul t to decide exactly at which point of
this process the state enterprise turns into a private firm. In most
cases, for a long time, the ownership remains mixed (de jure, or
de facto, or both). Examples of such mixed companies would
include:
Hidden employment and hidden unemployment. The official unemployment figures are an additional source of confusion. In general the following categories of unemployed workers can be identified:
Of course there are great many combinations of these categories. Some countries limit their unemployment figures only to those formally registered as unemployed. Apparently in some countries (for example in Russia) the unemployment allowance is so minus cule that many people do not bother to register in order to collect this benefit. In Poland, up to 50% of those registered as unemployed work in the unofficial sector. The expansion of this sector is also believed to help in keeping a low official unem ployment rate in the Czech Rep. In Albania about 50% of the labor force is out of work, although the official unemployment rate remains below 20%. The policies of soft budget constraint in the former Soviet republics kept millions at work despite a dram atic decline of output.
Output measurement. Another "calamity" of intertemporal and international comparisons is the exchange rate problem Since these rates do not reflect a true "utility" of national output, their use in GDP statistics results in GDP overestimation for count ries with an overvalued currency and in GDP underestimation for the countries with an undervalued currency. In addition, another serious problem of the exchange rate based statistics (GDP-X) is their volatility. During the last five years we witnessed dr amatic shifts in GDP-X which were produced by the shifts in exchange rates rather than the changes in real output. This latter problem can be illustrated by the following table which provides GDP-X figures (in billions of dollars) for selected Eastern Eu ropean countries, as published by EIU:
1991 1992 1993 -------------------------------------------------------------- Poland 63 Russia 69 Russia 167 Romania 38 Poland 59 Poland 73 Hungary 33 Hungary 36 Hungary 36 Czech. Rep. 32 Yugoslavia 28 Czech. Rep. 30 Yugoslavia 28 Belarus 24 Yugoslavia 28 Russia 28 Czech. Rep. 19 Romania 21 Belarus 3 Romania 18 Belarus 5
In light of these statistics, between 1991 and 1993 the GDP of Russia increased by a factor of six; between 1991 and 1992 the GDP of Romania decreased by more than half, while that of Belarus increased by eight times.
An alternative to the GDP-X statistics are the GDP-PPP statistics (purchasing power parity based) which are assumed to reflect changes in a "real" domestic product. They are, however, difficult to calculate and suffer from a high arbitrariness. As illus trated in the following tables they differ significantly from source to source and therefore their reliability (and usefulness) is quite limited:
GDP-PPP, Soviet Republics, 1991, thousands of dollars per capita
PENN PlanEcon World Bank EIU
Russia 8.1 Estonia 8.5 Belarus 7.5 Estonia 6.7
Latvia 6.7 Latvia 8.2 Russia 7.5 Latvia 6.1
Estonia 6.2 Belarus 7.5 Estonia 6.4 Belarus 5.5
Belarus 5.3 Russia 7.2 Latvia 7.1 Russia 5.2
Sources:
PENN: Data extracted from the Penn World Tables (Mark 5.6); for
Belarus - estimated from the 1990 data
PlanEcon: PlanEcon Report, March 27, 1992
World Bank, World Development Report, 1994, estimated from the 1992
data
EIU: the Economist Intelligence Unit, different 1993-94 publications
GDP-PPP, East-Central Europe, 1991, thousands of dollars per capita
PENN PlanEcon ALTON Bulgaria 6.7 Czech-Sl. 7.4 Czech-Sl. 6.8 Hungary 6.0 Hungary 5.5 Hungary 5.8 Poland 4.6 Bulgaria 5.2 Poland 4.2 Czech-Sl. 4.5 Poland 4.1 Bulgaria 4.1Sources:
As shown in the tables above, according to PENN, the Russian GDP-PPP per capita was the highest among the four Soviet republics included in the table ($8,100), but according to EIU it was the lowest ($5,200). Bulgaria had the highest GDP according to PEN N ($6,700) and the lowest according to ALTON ($4,100). At the same time, Czechoslovakia had the highest GDP according to PlanEcon ($7,400) and the lowest according to PENN ($4,500). It should be noted that the relatively high levels of GDP-PPP for Russia and Belarus during the early 1990s (as recorded by almost all sources) remained in sharp contrast with the relatively low levels of household consumption in these countries (for exam ple, see Comparing, 1993, p. 10)
The good news is that, despite of all these "particular" problems, the correlation coefficients between GDP-X per capita and GDP-PPP per capita tend to be quite high (which is also the case of the data presented herein--see below).
Another fundamental problem with GDP measurement in communist and post-communist economies is the utility of product. The question remains how much of it is useful, how much of it is "socialist output" (a term coined by Leszek Balcerowicz), and how much of it is an accounting phenomenon and in reality is not produced at all. The following table provides a breakdown of Russian GDP, for 1991-1994:
1991 1992 1993 1994
Individual consumption 42.1 33.5 42.1 44.2
Government's current account expenditures
- Collective consumption 6.8 7.2 12.1
- Subsidies to enterprises 9.3 7.1 4.7
- Other 3.5 1.8 1.2
Fixed Investment 24.4 19.2 22.4 20.8
Inventories 13.6 15.5 4.4 2.8
Net exports 0.3 15.7 13.1 7.7
Total GDP 100.0 100.0 100.0 100.0
Sources: Voprosy Ekonomiki, 1995, #1, p.19; Rossijskaya Federatsya v
Tsifrakh v 1992 (p.13) and v 1993 (p.101).
It is interesting that, despite an apparent major decline of Russian fixed capital investment and a process of dramatic decapitalization of different industries and enterprises, despite reports about deteriorating buildings and equipment, only about a hal f of Russian GDP is spent on consumption, while the remaining half accounts for different forms of accumulation. These extremely high rates of accumulation are virtually unknown elsewhere in the world*. Are these rates real, or ar product of the Russian national accounting system?
The 1994 data tables
All problems discussed above are "embedded" in the statistics presented
herein. Therefore one must use caution in any application of these
figures. The tables provide information on 21 Eastern European
economies* , and are organized into four
regions:
- European countries of the Commonwealth of Independent States (CIS) -
including the Asian part of Russia
- the Baltic states
- East-Central Europe
- the Balkans
This area is inhabited by 358 million people: 148 million in Russia, 52
million in Ukraine, 39 million in Poland, 23 million in Romania, and 96
million in the remaining 17 countries, 1.5 - 10.5 million each.
Table 1 is for 1994 and includes the
following variables:
- Population
- GDP: GDP-PPP per capita (estimated from 1992 figures published in
the 1994 World Bank's World Development Report) and GDP-X per capita
(annual and average monthly)
- Average monthly wage (at dollar market exchange rates)
- Share of private sector
- Macroeconomic stability indicators: unemployment, inflation -
expressed by the Consumer Price Index (CPI), state budget balance
(deficit) and foreign debt.
Table 2 provides annual real GDP growth
rates, 1990-1995, calculated as simple arithmetic averages over a number
of estimates published by different data sources, in particular:
- Official publications of state statistical offices
- The Deutsche Bank Research Review
- The Economist Intelligence Unit
- European Bank for Reconstruction and Development
- International Monetary Fund
- JP Morgan
- The LINK Project
- OECD
- PlanEcon, Inc.
- The United Nations, Economic Commission for Europe
- The Vienna Institute
- Voprosy Ekonomiki, 1995, #1
- The World Bank
While in most cases the figures for 1990-93 were measures of growth rates that actually took place, the 1994 figures were preliminary estimates, and the 1995 figures were predictions generated in 1994 or at the beginning of 1995. The two rightmost colum ns of Table 2 provide overall cumulative measures of the transition period -- 1989-1994 and 1989-1995, respectively. Set for 1989 at 100, in 1995 GDP is predicted to amount to about 86 in East-Central Europe, 63 in Balkan countries, 58 in Baltic countrie s, and 41 in the CIS*.
It is interesting to notice that while 1990 was the worst year for Poland (the greatest decline of GDP), 1991 was the worst for the remaining East Central countries as well as for Romania, Bulgaria, Albania, and Croatia. In turn, 1992 was the worst for all former Soviet republics except Belarus and Ukraine. 1993 was the worst for Macedonia and Yugoslavia where transition reforms were delayed by the war. Belarus and Ukraine postponed radical reform until 1994. During the last three years, the best per formers were Albania and Poland1*. In 1995 all the countries are expected to improve their rates of growth. GDP will decline in only five countries: Russia, Belarus, Ukraine, Azerbaijan, and Macedonia.
To measure relationships among the variables several simple regression models were run. The results are presented in Figures 1-4 in the format of scatter diagrams that include a relevant regression line and regression equation (with the corresponding t-s tatistics provided in parentheses under the estimates of regression coefficients).
Figure 1 examines the relationship between GDP-X and GDP-PPP for the 21 countries. It was encouraging to find out that this relationship was in 1994 quite strong and the slope coefficient highly significant (t=7.67). For Belarus and the Czech Republic the GDP-PPP estimates remain on the high side while for Hungary they remain on the low side.
Figure 2 demonstrates that, as expected, the monthly GDP-X figures are strongly related to the average dollar monthly wages with Poland located above the regression line (actual wages higher than predicted wages) and the Czech Rep. and Hungary located below this line (actual wages lower than those predicted by the model). Slovenia is the unquestionable leader in Eastern Europe. Its GDP-PPP/capita, GDP-X/capita, and dollar wages are much higher than those in any other Eastern European country.
It is interesting to find out that in transition economies there occurs a trade-off between inflation and unemployment (Figure 3), similar to that described by the so-called Phillips curve (well known from macroeconomics textbooks). In transition economi es a high inflation rate is a price that must be paid for a low unemployment rate. As shown in Figure 4, however, the underlying relationships behind this tradeoff are somewhat different from those postulated by Phillips.
If the marginal labor productivity is positive then, ceteris paribus, higher employment (i.e., a lower unemployment) should result in greater output. This does not seem to hold in Eastern Europe where, paradoxically, an increase in unemployment rates bri ngs about an increase in output. A high official unemployment rate often denotes an imposition of a harder budget constraint. This improves work discipline and reduces hidden unemployment. As a result, labor productivity increases *.
High inflation is a result of an increase in money supply. The money is used by the state to finance a budget deficit that is generated by subsidies that go to enterprises in order to help them finance the hidden unemployment.
The mood among forecasters is good. In 1995 and 1996 inflation rates and budget deficits should shrink dramatically. Privatization will continue. Many Eastern European economies should expand at Asian tiger rates. Even for those few countries which ec onomies are still contracting the news is not bad: the growth is believed to await them just around the corner.
References
"Comparing East European Purchasing Power", 1993. Transition,
July-August 1993.
Kornai, Janos, 1992. The Socialist System. Princeton, NJ: Princeton
University Press.