Journal Issue
Share
Article

Russian Federation: Selected Issues

Author(s):
International Monetary Fund
Published Date:
December 2006
Share
  • ShareShare
Show Summary Details

II. What Explains Weakening Growth Linkages Between Russia and Other Countries in the Region?32

Since the dissolution of the former USSR, the 15 countries that emerged from it have undergone deep economic, social, and political transformations. Each country has introduced its own currency, seven have joined the World Trade Organization (WTO), and the three Baltic countries have become members of the European Union (EU). Several countries have attained market economy status, while the remaining countries have made significant progress in their transition to a market economy. In this process, the structures of these economies have changed so that the strong interconnections that characterized these economies at the start of their independence have weakened, while links to the rest of the world have strengthened.

While one would expect this reorientation of economic links to happen gradually over time, there are indications that the Russian financial crisis in 1998 precipitated this process. For example, simple correlations between real GDP growth in Russia, on the one hand, and in the other former Soviet Union (FSU) countries, on the other, dropped significantly following the Russian crisis (Table 1). In a recent IMF working paper, Shiells, Pani, and Jafarov (2005) (henceforth SPJ) investigate whether this drop is evident after controlling for other factors that have been claimed in the previous literature to be determinants of growth in the transition countries.33 They conclude that Russian growth was a significant determinant of regional growth before the Russian crisis, but this link weakened significantly following the crisis.

Results of reestimations of the set of equations presented in SPJ using the latest (and slightly revised) data available are similar to those in the original paper. While there are a few changes in the coefficients, the main findings of the original paper remain robust. Specifically, the results suggest that the growth linkages between Russia and the other FSU countries weakened significantly following the Russian crisis (Appendix I).

This chapter tries to explain what might have weakened the linkages between growth in Russia and growth in other countries in the region.34Sections AD consider possible transmission channels, including trade linkages, capital flows, and labor flows. Section E presents the results of a regression analysis built on the work presented in SPJ. Section F concludes.

Table 1.Simple Correlation Coefficients Between Real GDP growth in Russia and the Other FSU Countries, 1993–2004
1993–971998–20041993–981999–2004
Armenia0.270.060.26-0.35
Azerbaijan0.88-0.060.730.52
Belarus0.81-0.220.730.28
Estonia0.730.290.730.23
Georgia0.730.230.72-0.19
Kazakhstan0.880.720.86-0.11
Kyrgyz Republic0.930.530.910.53
Latvia0.530.410.510.09
Lithuania0.84-0.220.78-0.09
Moldova0.720.610.71-0.28
Tajikistan0.930.440.74-0.04
Ukraine0.980.620.820.07
Uzbekistan0.950.300.87-0.05
Average0.780.290.720.05
Sources: IMF, World Economic Outlook database; and Fund staff estimates.
Sources: IMF, World Economic Outlook database; and Fund staff estimates.

A. Possible Transmission Channels

The Russian crisis caused significant disruptions in trade and financial flows in the FSU area, which reduced growth in the whole region. In Russia, real GDP declined by more than 5 percent in 1998 after a 1.4 percent increase in 1997. Average growth in the other FSU countries declined to 2.7 percent in 1998–99 from 6.1 percent in 1997. These disruptions, however, cannot explain the drop in the correlation between Russian growth and growth in other FSU countries after the Russian crisis. Interestingly, this correlation remained low even after growth picked up again across all these countries, including Russia, suggesting that recent high growth in the other FSU countries was not driven by growth in Russia.

It is possible that the Russian crisis triggered changes in patterns of trade, investments, and labor flows, and thus had a knock-on impact in the FSU countries. For example, exports of these countries to the rest of the world surged shortly after the Russian crisis. There are also signs of improvements in efficiency of production (see Sections BD). Accordingly, the rest of the paper focuses on flows of trade, capital, and labor to and from FSU countries, as well as on other factors that enhanced the supply response to positive shocks these economies experienced.

Flows of trade, capital, and labor can affect growth both in the short and long term. While the literature on growth focuses mainly on determinants of long-term growth, short-term determinants of growth are discussed mainly in the context of transmission of business cycles. This paper takes into account both the short- and long-term impact of flows of trade, capital, and labor.

A number of theoretical models suggest that trade and financial openness promote long-term growth through knowledge spillover, development of the financial sector, augmentation of domestic savings, higher productivity, and investment growth. 35 Many empirical papers support these claims (Edwards, 1992; Dollar, 1992; and Sachs and Warner, 1995). The robustness of the findings of these empirical papers, however, has been criticized. Levine and Renelt (1992), for example, explain that, since policies correlated with growth (trade openness, macroeconomic stability, small government consumption, rule of law, etc.) are highly correlated among themselves, it can be difficult to identify separate effects of any of these policies, including trade openness, when all these policies are included in regression analyses. Rodriguez and Rodrik (2001) make a similar claim, suggesting that the findings of the rapidly increasing literature on the relationship between trade openness and growth are not robust to different specifications, and that the openness measures used in these studies may be capturing other policy and institutional features.

Regarding foreign financing, there is general agreement that long-term investments can enhance growth through various channels. First, long-term investments encourage the incorporation of new inputs and foreign technologies in the recipient economy. Some forms of investment, in particular foreign direct investment (FDI), usually affect growth directly by increasing the stock of physical capital. Second, they facilitate knowledge transfers through labor training and skill acquisition, as well as alternative management practices and organizational arrangements. As a result, foreign investments may increase productivity in the recipient economy and become a catalyst for domestic investment and technological development, thus enhancing growth permanently. The extent to which foreign investments are growth enhancing depends on the degree of complementarity and substitution of foreign investment and domestic investment (de Mello, 1999).

As for the impact of trade and capital flows on business cycle transmission, the existing theoretical literature does not provide definitive guidance. International trade could produce both demand- and supply-side spillovers across countries. While demand-side spillovers usually increase the degree of business cycle synchronization (e.g. greater output in trading-partner countries would raise net exports and thus growth in a given country), the impact of supply-side spillovers depends on the specialization patterns. For example, trade flows could increase specialization of production, which would weaken business cycle correlations. Financial linkages, in turn, usually increase business cycle correlations by strengthening comovements of consumption across countries. However, it is also possible that international financial linkages could stimulate specialization of production, which could result in more exposure to asymmetric shocks and thus reduce correlation in business cycles (Kose, Prasad, and Terrones, 2003).

Views on the impact of labor outflows and related inflows of remittances on growth and business cycle correlations differ widely. Some analysts argue that remittances help increase savings and investments (e.g., in real estate or starting up small businesses), allow for increased expenditures on education and health, raise output through the multiplier effect, and lead to “brain gain.” Therefore, they assert that remittances can help improve a country’s development prospects, maintain macroeconomic stability, mitigate the impact of adverse shocks, and reduce poverty (Adelman, Taylor, and Vogel, 1988; Meyer and Brown, 1999; and IMF, 2005). On the contrary, other researchers argue that remittances could weaken growth prospects in the recipient country due to reduced incentives to work (Addleton, 1992; and Chami, Fullenkamp, and Jahjah, 2003), appreciation of the recipient country’s currency because of “Dutch disease”-type effects of remittances (Bourdet and Falck, 2003), and “brain drain” (Desai, Kapur, and McHale, 2001). In any case, it would be safe to conclude that remittances can support the development process if the economic environment in the recipient countries is conducive to growth.

To conclude, there is no definitive guidance on how and to what extent the above factors affect growth, especially in the medium term (or over the business cycle). Ultimately, an empirical analysis is needed to estimate the impact of each factor in a given country or region. In sections B, C, and D we analyze the role respectively of trade, financing, and labor-associated flows in descriptive terms. In section E, the analysis is conducted with econometric tools.

B. Foreign Trade Linkages

With the disintegration of the former USSR and transition to a market economy, FSU countries reoriented their trade away from the FSU area, including Russia, and toward markets in western countries.36 Most of the decline in the share of trade with Russia in the total trade of the other FSU countries was due to declining shares of exports to Russia (in total exports). The drop in this ratio was especially pronounced for the Baltic countries, which were pursuing EU membership. The shares of Russia in the imports of the other FSU countries declined less and more gradually (Figure 1).

Figure 1.FSU Countries: Selected Economic Indicators, 1993–2004

Sources: IMF, International Financial Statistics, Directions of Trade Statistics, and World Economic Outlook databases; and Fund staff estimations.

The Russian crisis accelerated the trade diversification of the FSU countries This outcome was mainly related to reduced demand in the region, due to lower growth, and to significant changes in the competitiveness of the FSU countries vis-à-vis Russia and the rest of the world after the crisis. In 1998–2001, the currencies of the countries of the Commonwealth of Independent States (CIS) appreciated against the Russian ruble but depreciated against the U.S. dollar and other convertible currencies.37, 38 Accordingly, these countries lost in competitiveness against Russia, but gained in competitiveness against other countries. The three Baltic countries that were pursuing memberships in the EU defended their pegs against hard currencies despite the significant adverse impacts on their external balance of the Russian crisis.

The large devaluations significantly improved the current account balances of the CIS countries. Imports declined significantly due to both (i) import substitution, caused by the devaluations; and (ii) declines in overall demand, caused by the “wealth effects” of the devaluations and reduced growth. As a result, imports in 1999 on average fell by more than 3 percent of GDP, and the current account balances improved significantly.39 While growth in imports picked up after 1999, it was offset by a surge in exports.40

Table 2.FSU Countries: Merchandise Exports as a Percent of GDP, 1993–2004
1993199419951996199719981999200020012002200320041994–971998–992000–04
Armenia20.537.127.818.214.211.612.615.716.121.224.716.724.312.118.9
Azerbaijan75.928.222.619.919.713.720.433.240.634.825.628.022.617.032.4
Belarus52.250.744.139.251.546.748.957.460.054.755.860.146.447.857.6
Estonia49.254.449.044.759.358.554.269.967.261.761.251.651.956.362.3
Georgia29.718.98.06.56.49.112.910.79.910.311.714.210.011.011.4
Kazakhstan27.731.728.429.425.533.054.041.039.341.951.129.329.345.5
Kyrgyz Republic54.030.632.427.934.631.236.636.831.230.230.331.831.433.932.1
Latvia47.927.126.225.527.227.423.924.124.825.126.129.226.525.625.9
Lithuania41.746.642.340.639.233.427.733.537.939.039.041.842.230.638.2
Moldova35.939.251.846.945.337.339.636.638.438.739.938.045.838.438.3
Tajikistan51.659.3142.074.071.445.263.477.761.760.850.944.186.754.359.0
Ukraine40.340.832.328.430.236.746.642.842.348.450.235.433.446.1
Uzbekistan30.526.718.819.715.511.515.917.916.222.626.023.913.519.7
CIS, excl. Russia36.236.630.829.827.931.242.839.638.541.745.433.429.541.6
Baltics45.641.638.837.140.537.732.838.740.540.040.640.539.535.240.1
Sources: IMF, Direction of Trade database; and Fund staff estimates.
Sources: IMF, Direction of Trade database; and Fund staff estimates.

The surge in exports of the FSU countries was mainly due to increased exports to the rest of the world (excluding Russia). This growth was related to gains in competitiveness following sizable devaluations (in the CIS countries) and improvements in the terms of trade, including higher oil prices for oil-exporting countries, after the Russian crisis. Several other factors enhanced the supply responses to the positive shocks these countries faced, including (i) ample idle resources; (ii) the imposition of harder budget constraints; (iii) improvements in financial discipline; (iv) achievement of macroeconomic stability; and (v) accumulated structural reforms.

  • A collapse of output during the initial years of the transition generated ample idle resources in all the FSU countries. For example, estimates by the Institute for the Economy in Transition (IET), the Russian Economic Barometer (REB, 2004), and the Center for Economic Analysis (CEA) all suggest significant declines in capacity utilization in Russia from 1993 to 1998 (see also Oomes and Dynnikova, 2006). Berengaut and others (2002) provide evidence of idle capacities in Ukraine. Anecdotal evidence suggests that such idle resources were abundant in other FSU countries as well. When demand picked up after the Russian crisis, these idle capacities helped increase output with minimal investment.
  • The Russian crisis heightened the sense of urgency among policymakers in all of the FSU countries to reduce loss-making activities. The heavier debt burden, related to the devaluations, together with foreign lenders’ desire to reduce their exposure to the FSU countries, forced the governments to tighten fiscal policies. Tighter fiscal policies, in turn, facilitated the imposition of harder budget constraints on loss-making activities and increasing the efficiency of production.
    • General government budget balances improved in all CIS countries with the exception of Belarus. In Russia, the government was able to produce a “remarkable fiscal adjustment at the general government level” amounting to 10 percent of GDP from 1997 to 2001 (Owen and Robinson, 2003).
    • Reported budgetary subsidies were (further) reduced in Azerbaijan, Kazakhstan, Lithuania, Ukraine, and Uzbekistan (EBRD Transition Reports, various issues).
    • There are indications that in many FSU countries a tighter financial situation caused further restructuring at the enterprise level. For example, there were significant declines in employment and jumps in unemployment rates in a number of FSU countries: employment declined in Georgia, Lithuania, Moldova, Tajikistan, Kazakhstan, and Ukraine; unemployment rates rose in Armenia, Georgia, and Ukraine. In addition, the sectoral distribution of labor resources changed significantly, with the share of the industrial sector in total employment declining and the shares of the agriculture and service sectors increasing.41
    • Country case studies also indicate that budget constraints hardened in some countries. Berengaut and others (2002), for example, suggest that after the Russian crisis budget constraints were hardened in Ukraine.
  • Better payment discipline in the aftermath of the Russian crisis improved the business climate. Chains of arrears were broken due to pressure from the governments, better enforcement of the existing rules against nonpayments, and greater liquidity in the financial systems, which was in part related to the increases in export revenues. Barter, payment arrears, wage arrears, and tax arrears declined significantly, which accelerated growth in economic activity, including investments.42
  • These factors also helped achieve macroeconomic stability, which shifted the focus of producers away from inflation-hedging activities toward productive activities. In particular, hardening the budget constraints on state-owned enterprises and narrowing the budget deficits allowed inflationary financing from central banks and inflation to be reduced.
  • Finally, the FSU countries benefited from the accumulated reforms they had undertaken prior to the Russian crisis. These reforms obviously were not enough to produce sustainable growth before the crisis, in part due to overvalued exchange rates, but were enough to enhance the supply response to positive impulses to the economy after the crisis. Havrylyshyn and De Souza (forthcoming), for example, consider the levels of reforms achieved by the Central and Eastern European countries before growth resumed in these countries (measured by the EBRD transition index) as the threshold needed to stimulate local economic activity. They suggest that Kazakhstan, Georgia, and Armenia reached the threshold level of reforms in 1996–97, just before the Russian crisis, and Ukraine, Azerbaijan, and Tajikistan reached the threshold level in 2000–03.

It is also possible that the Russian crisis forced exporters in the other FSU countries to incur significant “sunk entry costs” to access global markets.43 These exporters redirected their exports from Russia toward other markets as they lost in competitiveness to their competitors in Russia and demand in Russia declined during the crisis. In this process, these enterprises may have gained new connections and marketing knowledge. For example, after 1999, the Baltic countries’ exports to the rest of the world continued to rise rapidly despite the significant real appreciation of their currencies.44

Baltic Countries: Real Effective Exchange Rate and Trade Vis-à-vis Non-FSU Countries

(1997=100; millions of 1997 dollars)

Sources: IMF, International Financial Statistics and Directions of Trade Statistics databases; and Fund staff estimations.

Exports to Russia (in U.S. dollar terms), however, declined significantly after the Russian crisis. In 11 out of the 13 countries, exports to Russia did not reach the 1997 levels before 2003, even though the Russian ruble appreciated in real terms against the currencies of many FSU countries in 2001–03 (Figure 1). In 2004, only Belarus, Kazakhstan, and Ukraine recorded significantly more exports to Russia than the 1997 levels, while 7 out of the 13 countries recorded less exports than the 1997 levels.

The decline in exports to Russia was in part related to Russia’s tightening of the terms of payments for delivered goods and services and curtailing of its financing of other FSU countries.45 Before the Russian crisis, the FSU countries paid in-kind for a large share of their imports (mainly energy) from Russia, which increased these countries’ exports to Russia. Russia was also generally lenient on nonpayments for delivered goods. After the crisis, Russia demanded cash payments for its exports, which led to declines in its imports from the FSU countries. For example, in 1997, the shares of barter in export and import transactions between Russia and the CIS countries were about 20 percent and 24 percent, respectively. Following the Russian crisis, these ratios declined, reaching 1 percent and 5 percent, respectively, in 2004 (Table 3).

Table 3.Share of Barter in Export and Import Transactions of CIS Countries, 1993–2004(In millions of U.S. dollars; unless otherwise indicated)
199319941995199619971998199920002001200220032004
Export, f.o.b. based14,88014,12414,61415,89516,62413,69910,70713,80114,61715,71120,54029,462
of which: barter1,4352,7033,6393,2872,8961,9931,7031,239898519323
Share of barter in total export transactions (in percent)10.218.522.919.821.118.612.38.55.72.51.1
Imports9,69910,31713,59214,54914,23411,3148,36111,61011,20210,16313,13917,721
of which: barter2,3463,3363,8683,4173,1822,0041,9191,7531,160904847
Share of barter in total import transactions (in percent)22.724.526.624.028.124.016.515.611.46.94.8
Source: Central Bank of Russia.
Source: Central Bank of Russia.

Exports of other FSU countries to Russia may also have been adversely affected by the reduction in energy subsidies from Russia to other FSU countries. Russia has provided some FSU countries with subsidies in the form of cheap prices for energy products, mainly gas, but for some time now it has been trying to reduce these subsidies. To the extent that these subsidies are reduced far more for producers in the other FSU countries than for Russian producers, the price increases for imported energy in the FSU countries reduced the competitiveness of producers in these countries against Russian producers.46

Growth decompositions suggest that the contribution made by exports to Russia to growth in other FSU countries declined following the Russian crisis (Table 4).47 On average, exports (net exports) to Russia contributed almost 2.5 percentage points (1 percentage point) of GDP growth in the other FSU countries during 1994–97.48 This number declined to less than 1 percentage point (about -1 percentage point) after the Russian crisis. These figures should be interpreted with caution, however, due to weaknesses in data, including many missing observations for the early years of transition.

Table 4.Contribution of External Demand and Exports to Russia to Real GDP Growth in the CIS and Baltic Countries, 1993–2004 (In percentage points of real GDP)
Averages
1993199419951996199719981999200020012002200320041994–971998–20041999–20042000–04
ArmeniaReal GDP-14.15.46.95.93.37.33.36.09.613.213.910.15.49.19.310.6
Domestic demand-4.9-3.815.010.98.03.17.78.54.611.625.14.39.810.111.5
Net exports10.7-9.1-7.5-0.70.2-1.71.18.62.3-15.0-2.0-0.7-0.7-0.9
Exports14.014.7-4.4-3.30.50.24.94.411.16.1-9.65.22.52.83.4
o/w Russia5.10.10.4-2.3-1.7-0.70.81.50.21.1-1.6-0.6-0.10.20.4
Imports-3.7-4.0-4.7-4.3-1.20.0-6.5-3.3-2.5-3.8-5.3-4.2-3.2-3.6-4.3
o/w Russia1.16.32.1-6.41.5-0.22.4-3.1-0.41.11.70.80.40.20.3
AzerbaijanReal GDP-23.1-19.7-11.81.35.810.07.49.26.58.111.510.2-6.19.08.89.1
Domestic demand-36.4-28.110.016.313.38.77.50.2-9.05.227.7-9.57.76.76.3
Net exports16.716.3-8.7-10.5-3.3-1.31.76.317.16.3-17.53.41.32.12.8
Exports21.123.7-6.6-4.20.80.25.96.817.88.8-13.08.53.94.45.3
o/w Russia4.61.8-1.30.6-1.5-1.9-0.6-0.71.20.4-0.41.4-0.5-0.30.0
Imports-4.4-7.4-2.1-6.3-4.1-1.6-4.2-0.5-0.7-2.5-4.6-5.1-2.6-2.3-2.5
o/w Russia5.40.5-2.1-2.7-0.1-2.5-0.64.0-2.40.3-1.10.3-0.4-0.40.0
BelarusReal GDP-7.6-11.7-11.32.811.48.43.45.84.75.07.011.0-2.26.56.26.7
Domestic demand4.712.311.42.69.42.17.411.414.08.58.47.88.9
Net exports-2.0-0.9-3.00.7-3.62.6-2.4-4.4-3.0-1.4-1.9-1.7-2.2
Exports17.427.5-11.511.48.06.78.29.412.422.56.49.48.9
o/w Russia13.724.8-6.7-0.91.25.51.24.24.019.31.22.53.2
Imports-19.4-28.48.5-10.6-11.6-4.1-10.6-13.8-15.5-23.9-8.3-11.0-11.1
o/w Russia-6.9-17.33.6-7.3-14.5-3.4-6.5-9.7-13.0-12.1-7.2-9.1-9.4
EstoniaReal GDP-8.21.04.54.411.14.40.37.96.57.26.77.85.35.86.17.2
Domestic demand-8.65.45.27.613.05.8-4.68.58.59.911.28.47.86.87.09.3
Net exports0.4-4.4-0.7-3.2-1.9-1.35.0-0.7-2.1-2.6-4.6-0.6-2.5-1.0-0.9-2.1
Sum-8.21.04.54.411.14.40.37.96.57.26.77.85.35.86.17.2
Exports13.812.23.21.617.38.30.621.1-0.10.74.512.48.66.86.57.7
o/w Russia3.0-2.6-0.54.6-2.7-3.0-0.41.61.21.7-3.81.1-0.8-0.40.1
Imports-13.4-16.6-3.9-4.8-19.2-9.64.4-21.8-1.9-3.3-9.1-12.9-11.1-7.7-7.4-9.8
o/w Russia-2.5-0.31.0-3.41.5-1.4-2.80.60.00.6-0.3-1.3-0.3-0.5-0.4
GeorgiaReal GDP-10.42.610.510.62.93.01.94.75.511.16.23.35.05.45.9
Domestic demand11.9-5.5-2.811.85.07.311.65.411.94.76.48.2
Net exports-1.48.45.8-9.8-0.3-1.8-0.50.8-1.40.4-1.0-2.3
Exports2.69.41.9-3.70.32.35.26.22.63.12.02.1
o/w Russia1.0-0.7-0.71.50.9-1.41.10.51.00.20.30.5
Imports-4.0-0.93.9-6.1-0.6-4.0-5.7-5.4-4.0-2.7-3.0-4.4
o/w Russia0.91.51.0-3.70.5-1.9-0.3-0.70.9-0.5-0.8-1.2
KazakhstanReal GDP-9.2-12.6-8.30.51.6-1.92.79.813.59.89.39.4-4.77.59.110.4
Domestic demand-27.8-10.24.5-1.18.01.321.512.33.98.1-11.27.79.29.4
Net exports19.510.7-2.9-0.8-5.38.5-8.0-2.55.41.39.1-0.2-0.11.0
Exports22.715.90.9-0.7-6.714.4-3.21.011.26.813.23.33.96.1
o/w Russia7.60.6-0.6-2.7-2.53.50.4-1.31.02.82.60.20.71.3
Imports-3.1-5.2-3.8-0.11.4-5.9-4.8-3.5-5.8-5.5-4.1-3.5-4.0-5.1
o/w Russia-6.5-5.01.52.51.4-7.5-0.41.5-2.40.0-3.3-0.7-1.2-1.8
Kyrgyz Rep.Real GDP-13.0-19.8-5.87.19.92.13.75.45.30.07.07.1-2.14.44.85.0
Domestic demand-22.6-7.320.2-3.55.85.12.72.92.59.28.0-3.35.25.15.1
Net exports2.81.5-13.113.3-3.7-1.42.72.4-2.5-2.2-1.21.1-0.8-0.4-0.2
Exports-6.3-5.84.66.7-3.1-3.32.8-0.91.92.74.8-0.20.71.32.3
o/w Russia-5.90.82.1-2.2-0.4-0.8-0.30.11.10.51.6-1.30.20.40.6
Imports9.17.4-17.66.6-0.61.9-0.23.4-4.5-4.9-5.91.4-1.5-1.7-2.4
o/w Russia7.60.0-2.2-0.60.92.8-2.12.5-1.3-2.8-0.51.2-0.1-0.3-0.9
LatviaReal GDP-11.42.2-0.93.88.34.73.36.98.06.47.58.53.36.56.87.5
Domestic demand-13.96.64.07.86.111.93.24.511.46.212.413.46.19.08.59.6
Net exports2.5-4.4-4.9-4.02.2-7.20.12.4-3.40.3-5.0-4.9-2.8-2.5-1.8-2.1
Exports9.9-0.63.07.45.62.2-2.84.92.92.61.83.63.92.22.23.2
o/w Russia-0.3-0.31.10.2-3.7-2.6-0.80.80.2-0.10.60.2-0.8-0.30.2
Imports-7.3-3.8-7.9-11.5-3.4-9.42.9-2.5-6.3-2.4-6.8-8.5-6.7-4.7-3.9-5.3
o/w Russia0.4-1.1-1.81.80.81.0-0.90.60.0-0.5-0.8-0.20.0-0.1-0.3
LithuaniaReal GDP-16.2-9.83.34.77.07.3-1.73.96.46.89.76.71.35.65.36.7
Domestic demand10.78.5-0.32.25.96.812.112.710.76.86.67.9
Net exports-3.7-1.2-1.31.80.5-0.1-2.4-6.0-3.7-1.3-1.3-1.2
Exports8.62.4-8.44.29.510.04.02.48.63.43.66.0
o/w Russia2.4-3.7-5.40.32.81.8-0.7-0.32.4-0.7-0.20.8
Imports-12.3-3.67.1-2.4-9.1-10.0-6.4-8.4-12.3-4.7-4.9-7.2
o/w Russia-2.81.72.0-4.4-1.20.2-1.9-2.6-2.8-0.9-1.3-2.0
MoldovaReal GDP-1.2-30.9-15.3-5.91.6-6.5-3.42.16.17.86.67.3-12.62.94.46.0
Domestic demand23.5-38.4-9.58.810.8-0.5-24.518.47.212.927.419.2-7.18.610.117.0
Net exports-24.77.5-5.8-14.6-9.1-6.021.2-16.3-1.1-5.1-20.9-12.0-5.5-5.8-5.7-11.1
Sum-1.2-30.9-15.3-5.91.6-6.5-3.42.16.17.86.57.2-12.62.84.45.9
Exports-94.0-5.112.3-5.11.0-15.51.44.88.711.610.917.60.85.79.210.7
o/w Russia2.34.80.73.0-10.9-5.13.73.40.55.54.02.70.22.03.4
Imports69.412.6-18.1-9.6-10.19.419.7-21.2-9.8-16.7-31.8-29.6-6.3-11.4-14.9-21.8
o/w Russia1.2-0.5-0.8-1.87.93.52.5-2.2-1.2-2.4-2.5-0.50.8-0.4-1.2
TajikistanReal GDP-11.1-21.4-12.5-4.41.85.23.88.310.29.110.210.6-9.18.28.79.7
Domestic demand19.66.521.818.216.516.516.5
Net exports-9.42.6-11.6-7.6-6.5-6.5-6.5
Exports-9.17.2-4.9-0.7-1.9-1.9-1.9
o/w Russia-16.5-2.0-3.70.0-5.5-5.5-5.5
Imports-0.4-4.5-6.7-6.9-4.6-4.6-4.6
o/w Russia-3.2-4.00.5-1.4-2.0-2.0-2.0
UkraineReal GDP-14.2-22.9-12.2-10.0-3.0-1.9-0.25.99.25.29.612.1-12.05.76.98.4
Domestic demand-23.7-20.7-9.2-3.4-4.1-1.90.49.53.311.49.5-14.34.05.46.8
Net exports0.88.6-0.90.42.11.75.4-0.41.9-1.82.52.21.61.61.5
Sum-22.9-12.2-10.0-3.0-1.9-0.25.99.25.29.612.1-12.05.76.98.4
Exports-4.71.63.6-1.3-5.8-4.38.03.63.24.56.1-0.22.23.55.1
o/w Russia-1.90.51.0-6.2-2.8-1.93.20.2-1.50.81.2-1.7-0.10.30.8
Imports5.57.0-4.51.78.06.0-2.6-4.0-1.3-6.4-3.62.4-0.6-2.0-3.6
o/w Russia3.09.9-7.53.02.63.21.10.5-0.6-1.8-4.02.10.1-0.3-1.0
UzbekistanReal GDP-2.3-4.2-0.91.62.52.13.43.34.13.11.67.4-0.33.63.83.9
Domestic demand-6.26.51.55.15.55.05.910.71.36.40.65.75.85.9
Net exports5.3-4.91.0-3.0-2.0-1.7-1.8-7.60.31.00.4-2.1-2.0-2.0
Sum5.3-4.91.0-3.0-2.0-1.7-1.8-7.60.31.00.4-2.1-2.0-2.0
Exports8.1-1.14.2-2.9-2.80.60.0-6.03.14.43.7-0.5-0.10.4
o/w Russia-1.7-2.73.7-3.5-0.41.1-0.5-2.51.30.9-0.2-0.50.00.1
Imports-2.8-3.8-3.2-0.10.8-2.2-1.8-1.6-2.8-3.4-3.3-1.6-1.8-2.4
o/w Russia2.30.80.30.81.6-0.9-0.8-1.8-0.2-2.21.1-0.5-0.7-1.2
Average (excluding Uzbekistan, Tajikistan, Turkmenistan, and Russia)
GDP growth-11.8-11.7-4.42.36.13.32.05.97.36.89.18.8-1.86.06.47.3
Domestic demand0.3-16.3-11.06.18.14.9-0.36.87.55.811.613.80.47.07.48.8
Net exports-7.33.25.6-5.0-2.0-1.52.3-0.9-0.21.0-2.5-5.0-0.2-1.1-1.0-1.6
Total exports-23.54.49.43.85.6-1.2-0.96.83.56.46.34.56.43.34.15.1
Exports to Russia1.01.62.02.3-3.4-2.31.11.50.31.40.82.2-0.10.40.9
Total imports16.2-0.2-3.8-8.8-7.6-0.33.2-7.7-3.7-5.4-8.8-9.6-6.4-4.4-5.0-6.7
Imports from Russia2.31.0-2.6-2.52.20.3-2.8-0.2-1.2-1.8-2.2-1.1-0.8-1.2-1.6
Memorandum items:
RussiaReal GDP-8.7-12.7-4.1-3.61.4-5.36.310.05.14.77.37.2-4.85.16.86.9
Domestic demand-9.7-12.3-4.2-4.31.6-9.6-0.311.77.04.47.08.4-4.84.16.47.7
Net exports1.0-0.40.10.7-0.24.36.7-1.7-1.90.40.3-1.20.11.00.4-0.8
Sum-8.7-12.7-4.1-3.61.4-5.36.310.05.14.77.37.2-4.85.16.86.9
Exports-0.41.01.41.0-0.10.53.53.11.43.34.24.30.82.93.33.2
o/w FSU
Imports1.4-1.4-1.3-0.3-0.13.73.2-4.7-3.3-2.9-3.8-5.6-0.8-1.9-2.9-4.1
o/w FSU
Sources: IMF, Direction of Trade and World Economic Outlook databases; and Fund staff estimates.
Sources: IMF, Direction of Trade and World Economic Outlook databases; and Fund staff estimates.

A gravity model further suggests that the FSU countries reduced their regional trade and increased their trade with the rest of the world after the Russian crisis (see Jafarov and Svirydzenka, forthcoming).49 Nevertheless, the results also indicate that everything else given (e.g., levels of income and development, geographic location, etc.), among themselves, the FSU countries trade several times more than the estimates of the same model for non-FSU countries. The latter conclusion is consistent with the fact that the share of the regional trade in total trade of the FSU countries remains large, despite the large declines in this share. Russia, which has the largest economy in the region, remains a very important trading partner for the CIS countries. In 7 out of the 11 CIS countries, exports to Russia were more than 15 percent of total exports in 2004. In Belarus, this ratio was 47 percent.

C. Financial Flows

FSU economies have attracted large amounts of foreign savings since the beginning of their transition. Early in the transition, the bulk of foreign financing was from official sources to the public sector. As the transition progressed, the share of inflows from the private sector, including FDI and portfolio inflows, increased. Of these, FDI inflows were concentrated in the energy-rich Caspian countries and the Baltic countries. Weak FDI inflows to the other CIS countries, especially in the 1990s, reflected problems in the investment climate in these countries.50 The Baltic countries, Russia, and Ukraine attracted significant amounts of portfolio investment, but in the latter two countries these inflows dried up in 1998–99, following the Asian crisis and declines in oil prices.

Compared with the 1993–97 period, total foreign financing declined after the Russian crisis in all the FSU countries, except the three Baltic countries, mirroring the improvements in their current account balances (see Section B).51 The main factor contributing to this outcome was a decline in external borrowing. Meanwhile, FDI inflows in U.S. dollar terms declined during 1999–2000, but rebounded thereafter. During 2000–04, compared with the 1993–97 period, the ratio of FDI inflows to GDP increased in all the FSU countries with exceptions of the Kyrgyz Republic, Latvia, and Uzbekistan (Table 5).52 Portfolio investments rose in the Baltic countries but remained insignificant in the CIS countries.

Table 5.FSU Countries: FDI Inflows as a Percent of GDP, 1993–2004
Average for
1993199419951996199719981999200020012002200320041993–971998–992000–04
Armenia0.42.01.13.211.66.65.53.37.32.83.31.79.14.4
Azerbaijan1.011.420.827.326.018.513.114.732.655.053.415.122.333.8
Belarus0.50.20.10.72.51.33.70.90.81.71.00.70.82.51.0
Estonia9.49.05.43.25.410.35.47.19.04.110.19.36.57.97.9
Georgia0.01.00.31.86.66.12.25.02.53.88.77.41.94.25.5
Kazakhstan9.25.55.87.16.96.29.47.013.810.56.810.56.97.89.7
Kyrgyz Republic1.54.04.11.74.75.33.1-0.5-0.10.32.46.53.24.21.7
Latvia2.44.23.76.88.55.44.85.31.62.82.74.85.15.13.4
Lithuania1.93.68.34.53.33.75.21.03.52.76.43.3
Moldova1.01.24.61.44.14.53.210.69.97.03.03.12.53.86.7
Tajikistan1.41.53.82.42.71.81.92.40.00.00.013.12.41.93.1
Ukraine1.40.40.71.21.21.81.51.92.11.62.82.61.01.62.2
Uzbekistan0.91.11.00.41.10.90.70.50.70.40.70.90.90.80.7
Average for CIS, excl. Russia2.01.63.43.96.06.65.14.64.86.58.310.23.45.86.9
Average for Baltics5.96.64.54.05.88.04.95.34.84.04.65.85.46.54.9
Sources: IMF, World Economic Outlook database; and Fund staff estimates.
Sources: IMF, World Economic Outlook database; and Fund staff estimates.

Regarding financial flows from Russia to the other FSU countries, available data suggest that these flows also declined significantly after the Russian crisis. While there are no data available for individual countries, aggregate figures for financial flows from Russia to the CIS countries suggest that the CIS countries received significant amounts of financial inflows (from Russia) in the early years of their transition. Most of these inflows were trade credits and loans between governments. After the Russian crisis, these inflows almost halved. As for other types of financial inflows to the CIS countries from Russia, FDI picked up starting only in 2003, while portfolio investments never were significant, with the exception of 1995 (Table 6).

Table 6.Loans, Trade Credits, and Investments from Russia to CIS Countries 1/(In millions of U.S. dollars)
199319941995199619971998199920002001200220032004
Loans and trade credits-2,833-65-2,0871,386-622673224-117-778166-733-1,493
Total loans-1,633497-5451,04226844515383-246-43266260
Government (net)-1,633497-5451,042274848559386-695213723
Newly issued and restructured debt-1,670-216-1,172-13-1,330-47-1,245-79-313-448-80-254
Newly issued debt-1,670-216-37-13-67-47-150-60-400-175
Restructured00-1,1350-1,2630-1,229-79-253-408-80-79
Principle00-9690-1,1690-980-12-141-334-14-14
Interest rate00-1660-940-250-67-112-74-66-65
Repaid377136271,0551,6048951,804466244500217277
Central Bank00000000-51-950160
Banks217-48-4-116-353-12558
Nonbank entrprises-269-1131-10-365320
Trade credits-1,200-562-1,542344-648-171-291-500-532598-799-1,753
Direct investments-7-127-35-400-128-518-278-498-274-694-945
Portfolio investments-1,40081-5122392-43559175-31
Source: Central Bank of Russia.

Minus (plus) signs means an increase (decrease) in credits/investments from Russia.

Source: Central Bank of Russia.

Minus (plus) signs means an increase (decrease) in credits/investments from Russia.

Growth correlations between Russia and other FSU countries have likely been weakened by these foreign financing trends. Specifically, declining financial flows from Russia to the other FSU countries, at the same time when FDI to the FSU countries from other countries was increasing, may have loosened the relationship between growth in Russia and growth in the other FSU countries. Moreover, this dampening effect may have been exacerbated by the tighter payment terms and the reduction in implicit subsidies for Russian exports mentioned in Section B, which could be regarded as a form of financing to other FSU countries.

D. Labor Mobility and Associated Transfers in the Region

The FSU countries were characterized by extremely rigid labor markets in the early years of their transition. These rigidities were in large part due to legacies of socialism, such as little self-employment and entrepreneurship, the residential permit (propiska) system, the compressed administrative wage scale, the scarcity of part-time jobs, poorly targeted social assistance, sizable fringe benefits (housing, kindergartens, education, health services, etc.), rigid hiring and firing procedures, and in-kind payments. The misallocation of human resources under the planned economy, as well as transition-related factors, such as declines in output and weak enforcement of existing laws, exacerbated the negative impact of these rigidities.

Reducing these rigidities contributed to a more efficient allocation of labor resources in the economy.53 The Baltic countries pursued reforms vigorously, which initially led to rapid increases in unemployment. Subsequently, these economies’ capacity to create jobs expanded, and unemployment rates started to decline. In the CIS countries, a slower pace of reforms led to less dramatic increases in unemployment rates. Most of the adjustment came from large reductions in real wages and a buildup of large wage arrears during the 1990s. Poverty increased rapidly throughout the region (EBRD, 2000).

The lack of job opportunities at home and the possibility of earning more income abroad forced millions of people in FSU countries to migrate to other countries. Political tensions, wars, and increased levels of nationalism, as well as rapid population growth in some FSU countries, have further contributed to this trend. As a result, during 1991–2004, net emigration from the CIS countries, excluding Turkmenistan, Georgia, and Russia (data were not available for the first two countries) totaled more than 4.2 million.54 The structure of migration flows has also changed: while involuntary migration related to wars and regional conflicts dominated in the early 1990s, income motivated-emigration has been prevailing recently (Ivakhnyuk, 2003).

Russia has been a top destination for emigrants from many FSU countries. This outcome is due to cultural and linguistic ties, relatively low moving costs, the absence of a visa regime, the validity of diplomas and licenses issued in the other FSU countries, and the adverse demographic developments in Russia (Box 1). During 1993–2004, for example, accumulated (net) migration to Russia from Kazakhstan, Uzbekistan, and Ukraine was more than 2.5 million, according to Rosstat (Table 7).55 The actual numbers could have been much higher because a large share of labor inflows to Russia (estimated to be 2–5 million) are not reflected in official statistics. These significant labor force inflows suggest that labor markets in Russia or at least some segments thereof may have been considerably more flexible than previously thought (Box 2). For example, according to Rosstat, more than 43 percent of workers from the CIS countries in Russia were employed in the booming construction sector in 2004, suggesting that at least this segment of labor markets may have been rather flexible (Table 8).

Table 7.Migration between Russia and the Other FSU Countries
199319941995199619971998199920002001200220032004Total for
1993–2004
Migration to Russia
Total FSU982,5241,192,425869,081648,324593,755501,551370,602353,806188,802179,560123,433113,5226,117,385
Total CIS863,2481,100,273813,929614,022571,903488,087362,708346,774183,650175,068119,661110,3145,749,637
Azerbaijan54,68449,49543,44240,31029,87822,21015,90214,9065,5875,6354,2772,584288,910
Armenia2 9,80646,48034,11225,41919,12316,78014,67715,9515,8146,8025,1243,057223,145
Belarus34,67043,38335,33723,90317,57513,76011,54910,2746,5206,8415,3095,642214,763
Georgia69,93466,84751,41238,55124,51721,05919,62620,2139,6747,1285,5404,886339,387
Kazakhstan195,672346,363241,427172,860235,903209,880138,521124,90365,22655,70629,55240,1141,856,127
Kyrgyz96,81466,48927,80118,88613,75210,99710,37015,53610,74013,1396,9489,504300,976
Moldova19,34421,36418,71517,84713,75010,7629,03711,6527,5697,5626,3914,811148,804
Tajikistan68,76145,64541,79932,50823,05318,39612,11611,0436,7425,9675,3463,336274,712
Turkmenistan12,99020,18619,12922,84016,50110,5097,9986,7384,4024,5316,2993,734135,857
Uzbekistan91,164146,670112,31249,97039,62041,80041,61540,81024,87324,95121,45714,948650,190
Ukraine189,409247,351188,443170,928138,231111,93481,29774,74836,50336,80623,41817,6981,316,766
Baltic countries59,63846,07627,57617,15110,9266,7323,9473,5162,5762,2461,8861,604183,874
Estonia14,34011,2508,5915,8693,4831,77185278653553444544648,902
Latvia25,89126,37014,8598,2275,6583,5772,1081,7851,28399090681992,473
Lithuania19,4078,4564,1263,0551,7851,38498794575872253533942,499
Migration from Russia
Total FSU370,697232,810230,164192,205150,163133,567127,80782,31261,57052,09947,31037,9881,718,692
Total CIS362,941227,830225,876188,453146,961131,050127,80782,31261,57052,09946,08136,9501,689,930
Azerbaijan11,5436,1245,6144,9024,3023,9153,8473,1872,1701,7041,7711,33650,415
Armenia1,9531,9062,8402,9972,5782,3562,2431,5191,3621,1141,09865422,620
Belarus46,05827,75125,22921,54218,92819,03519,15113,27611,1758,8297,0165,663223,653
Georgia4,9224,6714,1094,1063,2862,9332,5741,8021,33996493974032,385
Kazakhstan68,70341,86450,38838,35025,36426,67225,03717,91315,18613,93914,01712,457349,890
Kyrgyz10,1429,9479,5518,4726,2965,3103,6811,8571,3331,08095965559,283
Moldova14,8819,3868,2646,8945,7154,7664,2752,2371,6601,3851,23490761,604
Tajikistan5,8983,6763,2902,6132,4741,9771,7991,15899382792254926,176
Turkmenistan6,1652,8171,9341,3801,5321,5371,23767635227225116818,321
Uzbekistan20,54511,31815,23513,3847,3705,2315,0413,0861,9741,4001,13071686,430
Ukraine172,131108,37099,42283,81369,11657,31858,92235,60124,02620,58516,74413,105759,153
Baltic countries6,1743,9223,4112,9302,5001,967878773
Estonia1,5821,058877822702550351265
Latvia2,2231,3391,167856636612259226
Lithuania2,3691,5251,3671,2521,162805268282
Net migration to Russia
FSU611,827959,615638,917456,119443,592367,98476,12375,534
CIS500,307872,443588,053425,569424,942357,037234,901264,462122,080122,96973,58073,3644,059,707
Baltics53,46442,15424,16514,2218,4264,7651,008831
Sources: Federal State Statistics Service (Rosstat); and Fund staff estimates.
Sources: Federal State Statistics Service (Rosstat); and Fund staff estimates.
Table 8.Russia: Distribution of Workers from the CIS Countries by Sectors of the Economy(In percent)
20002001200220032004
Industry17.217.817.816.717.8
Agriculture14.612.210.86.46.7
Transport7.97.07.511.38.0
Construction45.142.540.643.4
Other sectors20.017.921.325.024.2
Total100100100100100
Source: Rosstat; and Fund staff estimates.
Source: Rosstat; and Fund staff estimates.

Russia: Total Population Estimates 1/

(Millions)

Source: UN Population Database.

1/ Medium variant.

Russia: Inflow of Foreign Population

(In millions of people)

Source: Migration Information Source (MIS).

Data on transfers and remittances to the FSU countries, especially from Russia, are sketchy. Migrants often use unofficial channels to transfer money because of tax issues and the low credibility of financial institutions. For example, the IMF (2005) reports that elimination of the taxation of remittances in Tajikistan increased recorded remittances to $56 million in the first quarter of 2004 from $4 million in the first quarter of 2002.

Box 1.Russia: Demographic Developments, Labor Market Flexibility, and Immigration

Russia is suffering from a notable population decline, caused by low fertility and high mortality rates. The World Bank (2005a) reports that from 1992 to 2003 Russia’s population declined by about 6 million, mainly due to sharp increases in mortality and declines in fertility rates. While many developed countries experience low fertility rates, Russia’s mortality rate is high by comparison. This is mainly related to deaths from non-communicable diseases and injuries—specifically, heart disease, traffic accidents, and alcoholism—which account for 68 percent of deaths. If current trends continue, Russia’s population is expected to decline by over 30 percent during the next 50 years. Furthermore, the country’s population is aging rapidly, and significant numbers of people are emigrating. Enhancing internal migration and international immigration can, therefore, help reduce strains on Russia’s labor markets. In addition, migrants contribute to the development of their countries of destination by injecting social, cultural, and intellectual dynamism into these societies (GCIM, 2005).

Reportedly, net immigration into Russia surged from 130,000 persons a year during 1985–92 to a peak of 810,000 in 1994 and gradually declined since then (see also text figure on p. 15). In fact, in terms of both stock and flow of immigrants, Russia is second in the world only to the United States. Yet, many analysts believe that Russia needs more labor inflows. For example, Andrienko and Guriev (2005) estimate that, to compensate for the above demographic developments, Russia needs an annual inflow of 1 million immigrants—about 10 times the number officially recorded in recent years.

Recently, officially recorded immigration to Russia has slowed because of the winding down of inflows of ethnic Russians from the other FSU countries and enforced emigration, as well as the restrictions imposed on immigration. The introduction of the Law on Entry and Exit to/from the Russian Federation in 1996, as well as the amendments to it in 2000 and 2003, for example, raised the cost of residence permits. Many analysts argue that the present migration policy is too restrictive and, combined with high levels of corruption among government officials, forces otherwise legal immigrants into illegal immigration. Andrienko and Guriev (2005), for instance, suggest announcing an amnesty for the current illegal immigrants and introducing a point system to control the admission of new legal immigrants.

Box 2.Labor Mobility Within Russia

The very low interregional labor mobility in Russia (at about 1 percent)—despite substantial differences in wages and unemployment rates across regions—suggests that there are frictions in the labor markets. Explanations offered in the literature include the remnants of the Soviet style registry system (propiska); underdevelopment of the financial and property markets, which causes problems for people in selling and renting their houses; in-kind payments; and liquidity constraints.1 Kwon and Spilimbergo (2005) show that this situation was made worse by procyclical fiscal budgets in the regions.

The moderate levels of labor mobility between Russia and other countries, contrary to the low levels of interregional mobility within Russia, however, suggest that there can also be other explanations for regional disparities in unemployment rates. First, it is likely that elderly and less skilled labor, who would have difficulty finding jobs in any region, constitute a large share of unemployment. Anecdotal evidence suggests that many of these people resort to subsistence self-employment, primarily in agriculture, rather than moving to other areas.2 Second, there can be region-specific explanations as well. For example, very high rates of unemployment in the northern Caucasus and northern areas can in part be explained, respectively, by demographic trends in the northern Caucasus and privileges given to the people living in the north. The former is one of the few areas where the population continues to grow, due to its higher birth rates and life expectancy (World Bank, 2005b). Regarding the north, the current legislation requires that workers in this area be compensated with higher wages because of the arduous living conditions (World Bank, 2005a). It is possible that administratively determined higher wages reduce the competitiveness of enterprises in the north, causing (further) output declines and thus raising unemployment rates.

1See Andrienko and Guriev (2004), Friebel and Guriev (2005), and Andrienko and Guriev (2005).
2The share of immigrants in Russia working in the agriculture sector has been declining, perhaps due to lower wages in this sector (see Table 8).

Available data suggest that transfers and remittances (both from Russia and the rest of the world) constitute an important component of foreign financing for the FSU countries. On average, net transfers to the FSU countries (excluding Russia) rose from about 4 percent of their GDP during 1993–97 to 5 percent of GDP during 2000–04. During 2000–03, workers’ remittances, including workers’ compensation, amounted to about 4 percent of GDP. In 2003, remittances exceeded FDI inflows in Armenia, the Kyrgyz Republic, Moldova, and Tajikistan. In Moldova, for example, workers’ remittances totaled about 24 percent of GDP, compared with 3 percent of GDP in FDI inflows (Tables 9A and 9B).56 In Tajikistan, remittances were estimated at 14 percent of GDP in 2004 (Kireyev, 2006).

Table 9a.FSU Countries: Net Current Transfers-to-GDP Ratio(In percent)
1993199419951996199719981999200020012002200320041993–971998–992000–04
Armenia16.442.012.911.613.19.39.49.88.27.26.45.819.29.47.5
Azerbaijan, Rep. of2.33.34.62.11.11.41.81.41.31.11.82.12.71.61.6
Belarus1.40.50.70.60.60.60.91.21.21.41.21.20.80.71.3
Estonia6.44.73.42.22.42.72.02.12.51.61.31.53.82.31.8
Georgia15.920.79.74.55.76.36.56.87.15.46.46.511.36.46.4
Kazakhstan2.10.80.40.30.30.60.91.41.10.5-0.5-1.20.80.70.2
Kyrgyz Republic2.52.65.54.23.43.05.57.94.76.76.97.33.64.26.7
Latvia3.61.71.41.71.31.90.81.51.11.82.72.71.91.32.0
Lithuania0.00.00.01.82.32.11.52.12.11.71.61.50.81.81.8
Moldova1.41.63.94.32.83.97.412.916.015.115.413.62.85.614.6
Tajikistan-1.5-1.0-0.8-0.4-0.3-0.5-0.23.710.511.914.611.4-0.8-0.310.4
Ukraine0.90.80.51.11.71.92.22.73.84.54.44.01.02.13.9
Uzbekistan0.20.20.20.00.20.30.30.10.41.01.61.00.20.30.8
Average4.06.03.32.62.72.63.04.14.64.64.94.43.72.84.5
Average for the CIS countries4.27.23.82.82.92.73.54.85.45.55.85.24.23.15.3
Average for the Baltic countries2.21.61.92.02.21.41.91.91.71.91.92.21.81.9
Memorandum item:
Russia1.5-0.10.00.0-0.1-0.10.30.0-0.2-0.1-0.1-0.10.30.1-0.1
Sources: IMF, World Economic Outlook database; and Fund staff estimates.
Sources: IMF, World Economic Outlook database; and Fund staff estimates.
Table 9b.FSU Countries: Workers Remittances-to-GDP Ratio(In percent)
199319941995199619971998199920002001200220031993–971998–992000–03
Armenia5.15.38.34.85.14.64.45.56.16.25.05.1
Azerbaijan, Rep. of1.21.11.82.92.41.22.0
Belarus0.32.42.12.11.61.11.21.00.91.61.81.0
Estonia0.10.00.00.00.10.00.10.20.20.40.10.00.2
Georgia7.910.312.99.05.77.06.27.911.67.0
Kazakhstan0.40.30.30.40.70.70.80.50.7
Kyrgyz Republic0.30.10.10.10.20.10.73.71.84.25.60.10.43.8
Latvia0.70.70.70.71.72.51.51.50.70.71.8
Lithuania0.00.00.00.00.00.00.40.70.80.60.00.00.6
Moldova0.15.15.97.39.612.615.818.723.53.78.417.6
Tajikistan6.59.48.0
Ukraine0.00.00.00.10.10.40.50.70.4
Uzbekistan
Average0.30.10.91.62.62.62.93.23.24.14.82.63.34.0
Average for the CIS countries0.30.11.42.23.53.63.94.14.05.26.13.94.75.1
Average for the Baltic countries0.10.00.30.30.30.20.71.10.80.90.30.30.9
Memorandum item:
Russia1.60.80.70.60.70.70.50.50.40.30.90.70.4
Sources: World Bank, World Development Indicators database; and Fund staff estimates.
Sources: World Bank, World Development Indicators database; and Fund staff estimates.

Total net transfers from Russia to the other CIS countries were initially negative, but in recent years have turned positive and are growing rapidly. No data are available for individual FSU countries. However, aggregate figures for the CIS countries suggest that transfers from Russia to these countries were less than transfers from these countries to Russia for most of the 1990s. This was mainly due to migration-related transfers (migrants’ moving their financial assets from their home countries to their host countries) from the CIS countries to Russia, which peaked in 1994. Since then, these transfers have been declining, mirroring drops in the number of migrants from the CIS countries to Russia. In 2001, net transfers from Russia to the CIS countries turned positive as net remittances from Russia to the CIS countries exceeded migration-related transfers. From 2001 to 2004, net transfers from Russia to the CIS countries rose sixfold, due mainly to an eightfold increase in remittances (Table 10).

Table 10.Transfers and Remittances from/to Russia to/from CIS countries, 1993–2004(In millions of U.S. dollars)
199319941995199619971998199920002001200220032004
From Russia to CIS countries, net539-1,683439-467-133-329-263-3073936041,2942,916
Compensation of employees received002163093783031361483293276131,194
Remittances2024107121,785
Transfers related to migration539-1,683223-776-511-632-399-455-138-133-31-63
From CIS countries to Russia1,4283,9612,0362,6601,9831,578836752443446370616
Compensation of employees received000010000004
Remittances447593318
Transfers related to migration1,4283,9612,0362,6601,9821,578836752399371277294
From Russia to CIS countries1,9672,2782,4752,1931,8501,2495734458361,0501,6643,532
Compensation of employees received002163093793031361483293276131,198
Transfers of employees2464858052,103
Transfers related to migration1,9672,2782,2591,8841,471946437297261238246231
Memorandum items:
Migration to Russia from the CIS countries863,2481,100,273813,929614,022571,903488,087362,708346,774183,650175,068119,661110,314
Migration to the CIS countries from Russia362,941227,830225,876188,453146,961131,050127,80782,31261,57052,09946,08136,950
Total remittances to the CIS countries 1/269516729888
Sources: Central Bank of Russia; IMF, World Economic Outlook database; and Fund staff estimates.

Balance of payments data; Belarus, Turkmenistan, and Uzbekistan are excluded since data for these countries were not available.

Sources: Central Bank of Russia; IMF, World Economic Outlook database; and Fund staff estimates.

Balance of payments data; Belarus, Turkmenistan, and Uzbekistan are excluded since data for these countries were not available.

The recent increase in remittances from Russia to the CIS countries is in part related to wage growth in Russia, the increase in the number of illegal immigrants, booms in housing markets in the CIS countries, and weaknesses in data for the early years of transition:

  • Faster wage growth in Russia is likely to affect remittances in two ways. First, more income allows immigrants to increase their remittances, either for altruistic or investment purposes. Second, increases in the differences between wages in Russia and wages in the other CIS countries attract more labor inflows to Russia from these countries, including illegal immigrants.

Wages in the CIS Countries, 1993–2003

(In U.S. dollars)

Source: National authorities; and Fund staff estimates.

  • It can be argued that illegal immigrants tend to save and remit more than legal immigrants since illegal emigrants cannot integrate fully into their host countries and have limited options for investing in them.57
  • The increases in remittances to the CIS countries may also be related to housing market booms in these countries, as those migrants who want to invest in real estate in their home countries have to pay higher prices for these assets.
  • As discussed, the low remittance numbers reported in the earlier years of transition did not fully reflect actual flows (since many people avoided the official channels in transferring their savings).
  • It is difficult to measure the impact of the above factors on remittances, however, due to data shortcomings.

On balance, the weakening correlations between growth in Russia and growth in the FSU countries do not seem to be related to changes in remittances. In the CIS countries, the recent large and growing inflows of transfers and remittances from Russia may have strengthened domestic savings and investment, contributing to high growth in these countries (see Section A). These developments would suggest stronger rather than weaker correlations between growth in Russia and growth in the CIS countries.

E. Regression Analysis

In SPJ, growth in FSU countries was explained by initial conditions, growth abroad, and a number of variables to control for macroeconomic stability, progress in reforming, and trade openness. Two initial condition indices from Havrylyshyn and van Rooden (2000) were used to capture macroeconomic distortions and distortions related to the level of socialist development, of which only the second was statistically significant. Inflation and government expenditures as a percent of GDP (instead of the fiscal balance measure, which was not comparable across countries and had breaks in series) were included to measure the impact of macroeconomic stability. The EBRD transition index was used to measure the impact of reforms. The change in the relationship between growth in Russia and growth in the other FSU countries before and after the Russian crisis was measured by running piecewise regressions, where Russian growth was included as an explanatory variable and the coefficients on the explanatory variables were allowed to change. The latter was done by using the interactions of the dummy variable for the Russian crisis (0 before the crisis and 1 thereafter) with the explanatory variables. Both 1998 and 1999 were considered (separately) as a break point in the relationships, but it was eventually judged that the break point occurred in 1998 (see Appendix I).

This section enhances the analysis in SPJ by considering the role of some additional explanatory variables. Specifically, the FDI-to-GDP and transfers-to-GDP ratios were added to the specifications in Tables A1a and A1b to control for the impact of these two variables.58 In addition, this study employs the generalized least squares (GLS) method, adjusting the calculations using cross-section weights, instead of the least-squares dummy variable (LSDV) method.

The new variables improve the fit of the regressions (Table 11). This is reflected in the higher values for the coefficients of determination (R2) and/or log likelihood as well as the smaller values of the Akaike information criterion (AIC) for the regressions in Table 11, compared with those for the regressions in Tables A1a and A1b. The coefficients on transfers have the expected signs and are statistically significant (at the 5 percent level). The coefficient on transfers’ interaction with the Russian crisis dummy have minus signs, meaning that the impact of transfers on growth declined after the Russian crisis. The coefficients on FDI also have the expected signs and drop after the Russian crisis in all specifications except the second and third; however, FDI is significant only in the sixth specification.

Table 11.Coefficient Estimates in Real GDP Growth Regressions with Structural Break in 1998, CIS and Baltic Countries, 1993–2004 1/2/3/
Coef.t -stat.Coef.t -stat.Coef.t -stat.Coef.t -stat.Coef.t -stat.Coef.t -stat.Coef.t -stat.Coef.t -stat.
EQ1EQ2EQ3EQ6EQ7EQ8EQ9EQ10
C18.191.738.551.5523.003.6512.971.2518.511.655.321.1015.131.47-19.52-1.46
GR10.182.010.456.230.405.700.171.840.202.220.466.510.242.850.040.41
CPI0.00-1.900.001.190.000.660.00-2.300.00-1.270.001.200.00-2.36
INF-2.89-3.10
EXP-0.10-0.96-0.12-1.85-0.15-2.16-0.08-0.77-0.11-1.16-0.10-1.76-0.11-1.09-0.19-2.07
GRRUS0.905.260.554.260.815.480.985.350.804.370.534.200.643.700.674.21
EURGR-3.68-2.84-3.55-2.64-4.29-3.25-3.30-2.50-4.05-3.02-3.60-2.74-2.50-1.93
WORLDGR6.592.58
RER-0.06-3.15-0.03-1.94-0.04-2.67-0.09-2.23-0.03-1.82-0.05-2.45-0.05-2.54
RRUS-0.05-2.480.051.03
RI-2.41-0.763.142.55-1.32-0.78-2.07-0.64-1.29-0.393.933.53-3.50-1.15-0.86-0.28
OPEN-0.02-1.160.000.39-0.02-1.67-0.02-0.75-0.02-1.150.000.320.010.46-0.03-1.46
IC20.861.14
FDI0.301.780.050.530.030.280.352.060.241.410.080.830.211.360.241.51
TRS0.593.270.363.300.524.310.744.480.482.230.404.010.693.790.623.08
DGR10.100.71-0.09-0.64-0.01-0.060.130.900.090.61-0.03-0.190.030.180.221.53
DCPI-0.01-0.72-0.01-0.81-0.01-0.35-0.01-0.72-0.01-0.61-0.01-0.39-0.01-0.50
DINF0.870.34
DEXP-0.34-1.99-0.05-0.48-0.09-0.89-0.31-1.71-0.26-1.440.020.21-0.35-2.10-0.21-1.29
DGRRUS-0.83-4.27-0.39-2.53-0.65-3.85-0.90-4.39-0.74-3.64-0.35-2.36-0.58-2.94-0.67-3.46
DEURGR3.752.683.232.254.182.973.352.374.092.853.512.492.611.88
DWORLDGR-6.17-2.38
DRER0.071.580.010.150.010.370.131.990.000.030.051.210.061.41
DRRUS0.030.85-0.08-1.28
DRI6.841.16-4.60-2.831.170.554.700.754.410.70-4.10-2.807.851.356.451.21
DOPEN0.082.110.021.160.021.080.061.600.071.830.021.030.051.230.092.64
DIC20.540.52
DFDI-0.20-1.100.010.090.080.65-0.25-1.36-0.14-0.78-0.01-0.11-0.11-0.66-0.17-0.96
DTRS-0.31-1.29-0.31-2.04-0.41-2.58-0.50-2.15-0.22-0.83-0.37-2.54-0.41-1.61-0.35-1.38
GRRUS0.900.550.810.980.800.530.640.67
GRRUSxD98-0.83-0.39-0.65-0.90-0.74-0.35-0.58-0.67
Sum0.070.160.160.070.050.180.060.01
F -test0.543.803.990.640.354.700.480.01
P-value0.460.050.050.420.560.030.490.94
R20.900.850.870.900.900.850.900.91
Adjusted R20.850.820.840.850.850.820.850.86
Unweighted statistics
AIC2.72.92.82.72.73.02.72.8
R20.870.770.810.870.870.760.870.86
Source: Fund staff estimates.

The fourth and fifth specifications in Tables A1a and A1b, which estimate subsamples of the data set, are excluded from this table.

Bold indicates statistically significant at the 1 percent level; italics indicates significance at the 5 percent level.

Country/region fixed effects (for the first, third, sixth, seventh, ninth, and tenth specifications) are available upon request.

Source: Fund staff estimates.

The fourth and fifth specifications in Tables A1a and A1b, which estimate subsamples of the data set, are excluded from this table.

Bold indicates statistically significant at the 1 percent level; italics indicates significance at the 5 percent level.

Country/region fixed effects (for the first, third, sixth, seventh, ninth, and tenth specifications) are available upon request.

Controlling for FDI and transfers reduces the coefficients for Russian growth compared with those estimated in SPJ. However, Russian growth remains a statistically significant explanatory variable before the Russian crisis in all specifications, since FDI and transfers were not the only channels through which Russian growth affected growth in other countries. The coefficients on the interactions of the Russian crisis dummy with Russian growth have minus signs and are also statistically significant in all specifications.

The trade openness (OPEN) measure affects growth positively, but is not significant in many specifications. While the coefficient on trade openness is not significant in all specifications, the coefficient on its interaction with the Russian crisis dummy is significant in the first and tenth specifications and has the expected plus sign. This outcome might be related to the rather short sampling period, since trade openness usually affects growth in the medium and long term. Regarding the short-term impact of trade flows, the growth decomposition estimations in Section B suggest that trade flows were important factors of growth in the FSU countries.

The results suggest that, although Russian growth was a strong determinant of regional growth before the Russian crisis, this link weakened after the crisis. In the first specification (column 1), our benchmark, the coefficient on Russian growth is 0.9 while the coefficient on the interaction of Russian growth with the Russian crisis dummy is -0.8. This means that while 1 percent growth in Russia raised growth in other FSU countries by 0.9 percent before the Russian crisis, this effect declined to 0.1 percent after the crisis. Moreover, the sum of these two coefficients is statistically insignificant. The coefficient on growth in the EU has a minus sign and is statistically significant, which appears to be a reflection of declining output in the CIS countries when output in the EU was growing. The coefficient on the interaction of EU growth with the Russian crisis has a plus sign and is significant, suggesting that the EU took up the role that Russia used to play. Regarding other variables, the consumer price index (CPI), the reform index (RI), and the interactions of these two variables with the Russian dummy are not statistically significant. The coefficients on government expenditures (in percent of GDP) and trade openness are not significant, but the coefficients on their interactions with the Russian crisis dummy are significant and have, as expected, minus and plus signs, respectively. The coefficient on the real effective exchange rate (REER) has a minus sign and is significant. The coefficient on the interaction of REER and the Russian crisis dummy, however, has a plus sign and is not significant. 59

The conclusion that the growth linkages between Russia and other FSU countries weakened after the Russian crisis is robust to different specifications:

  • The results in the second specification—which includes an initial condition measure (IC2) but does not have country fixed effects—are broadly similar to those in the first specification. The coefficient for Russian growth was 0.6 before the crisis and dropped to 0.2 after the crisis. The sum of the two coefficients is not significant. The main differences from the results of the first specification are that (i) the coefficient on the interaction of government expenditures with the Russian crisis is not significant and (ii) the coefficients on the reform index and its interactions with the Russian crisis dummy have different signs and are significant. The sum of the latter two coefficients, however, is not significant. The coefficients of the initial conditions index and its interaction with the Russian crisis dummy are not significant.
  • The results of the third specification, which includes regional dummy variables for the Baltic, Caucasian, and Central Asian countries instead of dummies for each country, are similar to those in the first specification. The main difference from the results of the first specification is that the sum of the coefficients on Russian growth and its interaction with the Russian crisis dummy is significant in the third specification. The results of the sixth specification, which replaces the real effective exchange rate (REER) with the real exchange rate against the Russian ruble, as well as the results of the seventh specification, which includes both exchange rate variables (REER and the real exchange rate against the Russian ruble), are also similar to those in the first specification. 60
  • The results of the eighth specification, which excludes both country fixed effects and initial conditions, are similar to those in the second specification, which also excludes country dummy variables. The main difference here too is that the sum of the coefficients on Russian growth and its interaction with the Russian crisis dummy becomes significant.
  • The ninth specification, which tries to explore possible nonlinearity in the response of growth to inflation by including the natural logarithm of percent changes in CPI inflation (INF) instead of percent changes in CPI, also produces results similar to those in the first specification. The main difference is that the coefficients on EU growth and its interaction with the Russian crisis dummy become insignificant. Interestingly, the coefficient on (log of) inflation has a minus sign and is significant. The last specification, which includes world growth in place of EU growth, also produces similar results. Here, the coefficient on government expenditures has a minus sign and is significant at the 5 percent level.

The results should be interpreted with caution, however, because the right-side variables themselves, including Russian growth, could be endogenous and correlated among themselves. Moreover, it is well known that the inclusion of the lagged dependent variable in fixed-effect and random effect models creates biases in fixed-effect and random effect estimators. Therefore, the above results are compared with Arellano and Bond (1991) estimations, which use consistent instrumental variables. As can be seen in Table 12, the results from the Arellano-Bond estimations are similar to the results in Table 11, discussed above. In particular, the coefficients on Russian growth and their interactions with the Russian dummy variable are comparable to those in the equations presented in Table 11 and are significant. Arellano-Bond estimators, however, may exhibit a large bias in finite samples and will have larger standard errors than ordinary least squares. Therefore, the estimates presented in Table 12 should not necessarily be presumed superior to the GLS estimates presented in Table 11.

Table 12.CIS and Baltic Countries: Arellano-Bond Estimates of Real GDP Growth Regressions with a Structural Break in 1998, 1993–2004 1/
GLS EstimatesStrictly Exogenous Explanatory VariablesEndogeneity Correction 2/Endogeneity Correction 3/
(1)(11)(12)(13)(14)
Coefficientt -statisticCoefficientz-statisticCoefficientz-statisticCoefficientz-statisticCoefficientz-statistic
CONSTANT18.191.730.270.910.230.940.100.40-0.17-0.51
GR-10.182.010.273.910.385.760.385.710.575.28
CPI0.00-1.900.00-0.400.000.180.001.380.00-0.01
EXP-0.10-0.96-0.07-0.74-0.23-2.36-0.46-4.37-0.20-1.66
RI-2.41-0.761.880.63
GRRUS0.905.260.764.920.755.260.745.570.643.65
EURGR-3.68-2.84-3.38-2.62
WORLDGR0.070.130.200.370.100.16
RER-0.06-3.15-0.03-1.56
OPEN-0.02-1.160.00-0.12
FDI0.301.780.110.670.050.280.020.160.140.72
TRS0.593.270.473.160.533.590.513.590.552.83
D98-21.24-1.164.460.63-1.93-1.15-2.10-1.32-2.29-1.14
GR-1 ×D980.100.71-0.15-1.10
CPI×D98-0.01-0.72-0.01-0.74
EXP×D98-0.34-1.99-0.24-2.24
RI×D986.841.16-4.16-2.07
GRRUS×D98-0.83-4.27-0.71-4.41-0.62-3.53-0.63-3.87-0.43-1.88
EURGR×D983.752.683.252.41
RER×D980.071.580.020.37
OPEN×D980.082.110.062.05
FDIxD98-0.20-1.10-0.01-0.050.030.200.090.620.060.30
TRSxD98-0.31-1.29-0.09-0.40-0.23-0.98-0.23-1.23-0.01-0.03
Memorandum items:
Sargan test of overidentifying restrictions82.040.01 4/43.180.85 4/95.701.00 4/7.610.57 4/
Arellano-Bond test for AR(1)-2.610.01 4/-3.900.00 4/-3.600.00 4/-3.880.00 4/
Arellano-Bond test for AR(2)-0.640.52 4/0.420.67 4/0.280.78 4/0.510.61 4/
Source: Fund staff estimates.

Bold indicates statistically significant at the 1 percent level; italics indicates significance at the 5 percent level.

Coefficient estimates in this column assume that CPI and EXP are endogenous. First-order lagged values of these variables are used as instruments.

Coefficient estimates in this column assume that Russian growth and its interaction with the Russian crisis dummy is endogenous. First-order lagged values of these variables are used as instruments.

These figures refer to p-values instead of t-statistics.

Source: Fund staff estimates.

Bold indicates statistically significant at the 1 percent level; italics indicates significance at the 5 percent level.

Coefficient estimates in this column assume that CPI and EXP are endogenous. First-order lagged values of these variables are used as instruments.

Coefficient estimates in this column assume that Russian growth and its interaction with the Russian crisis dummy is endogenous. First-order lagged values of these variables are used as instruments.

These figures refer to p-values instead of t-statistics.

Finally, the results in Table 11 are compared with the ordinary least squares (OLS) estimations without country dummy variables (in Table 13) since small T (=12) may produce a bias in fixed-effect estimators as well as in Arellano-Bond estimations. The fit of the OLS regressions is uniformly worse than that in the regressions presented in Table 11, as reflected in the lower values for the R2 and/or log likelihood, as well as the higher values of the AIC for the OLS regressions. The main results, however, are similar to those above.

Table 13.Coefficient Estimates in Real GDP Growth Regressions with Structural Break in 1998, FSU Countries, 1993–2004 (OLS) 1/
EQ1EQ1EQ6EQ7EQ9EQ10
VariableCoef.t -stat.Coef.t -stat.Coef.t -stat.Coef.t -stat.Coef.t -stat.
C5.410.932.430.437.141.285.620.925.980.45
GR10.405.400.405.310.476.330.385.200.415.05
CPI0.001.220.001.130.001.76-0.15-0.140.001.14
EXP-0.09-1.24-0.07-0.96-0.10-1.54-0.08-1.19-0.11-1.62
GRRUS0.664.660.684.800.392.510.633.850.714.29
EURGR-3.48-2.15-3.53-2.17-4.54-2.88-3.24-1.93-2.14-0.72
RER-0.03-1.66-0.14-3.67-0.03-1.58-0.04-1.78
RRUS0.000.180.153.24
RI3.602.714.233.124.453.433.181.893.822.83
OPEN0.000.160.010.350.010.500.000.090.00-0.02
FDI0.070.520.120.900.090.660.060.460.040.28
TRS0.393.780.413.830.221.960.474.690.383.61
D98-0.13-0.022.200.32-1.75-0.26-0.24-0.03-2.21-0.16
DGR10.080.510.090.600.000.030.090.610.090.65
DCPI-0.01-0.62-0.01-0.64-0.01-0.72-1.19-0.35-0.01-0.66
DEXP0.000.00-0.01-0.070.030.31-0.01-0.070.030.32
DGRRUS-0.42-2.42-0.51-2.83-0.23-1.22-0.39-2.04-0.49-2.29
DEURGR3.151.833.472.004.442.642.901.622.250.74
DRER0.000.070.192.560.000.040.00-0.09
DRRUS-0.05-1.39-0.22-3.55
DRI-4.16-2.44-4.72-2.75-5.07-3.08-3.71-1.80-4.32-2.49
DOPEN0.020.840.010.560.010.490.020.890.020.92
DFDI-0.01-0.08-0.06-0.44-0.03-0.18-0.01-0.050.020.11
DTRS-0.41-2.65-0.42-2.68-0.24-1.53-0.49-3.16-0.40-2.49
R20.770.770.790.760.76
Adjusted R20.730.720.750.720.72
Log likelihood-393-393-385-394-395
Akaike information criterion5.805.815.725.825.84
F -statistic18.8718.7719.4618.5418.01
P -value (F -statistic)0.000.000.000.000.00
GRRUS0.660.680.390.630.71
GRRUSxD98-0.42-0.51-0.23-0.39-0.49
Sum0.240.180.160.240.22
F -test5.592.695.685.362.69
P -value0.020.100.020.020.10
Source: Fund staff estimates.

Bold indicates statistically significant at the 1 percent level; italics indicates significance at the 5 percent level.

Source: Fund staff estimates.

Bold indicates statistically significant at the 1 percent level; italics indicates significance at the 5 percent level.

F. Conclusions

Several transmission channels were considered as possible explanations for the weakening correlation between growth in Russia and growth in the other FSU countries following the Russian crisis. It appears that this weakening was related to changes in trade patterns and reduced capital flows from Russia to the other FSU countries. On the contrary, recent increases in remittances from Russia to the other FSU countries are expected to strengthen the growth linkages between Russia and other FSU countries in the future.

Following the Russian crisis, producers in the other FSU countries further diversified their trade away from Russia and toward markets in the rest of the world. The large devaluations in many FSU countries caused significant import substitution and boosted exports to the rest of the world. The latter process may have also strengthened the marketing skills and knowledge of exporters in the FSU countries. However, exports to Russia slowed, owing to the large devaluation of the Russian ruble in 1998–99 and reduced demand in Russia during the crisis. As a result, the contribution to growth in the FSU countries of exports to Russia declined from 2.2 percent in 1993–97 (the period of falling output in Russia) to less than 1 percent during 2000–04 (a high-growth period in Russia).

A number of other factors weakened the growth linkages between Russia and other FSU countries by enhancing the supply responses to positive shocks. For example, FSU countries built up sizable idle capacities due to the collapse of output during the first years of transition. When demand picked up, these capacities allowed output to increase with little investment. The supply responses were further boosted by the imposition of harder budget constraints, improvements in financial discipline, achievement of macroeconomic stability, positive impacts of the surge in export revenues, and accumulated structural reforms. Some of these factors, however, are temporary in nature. For example, excess capacities in the FSU countries are rapidly disappearing because of the surge in growth after the Russian crisis.

Capital flows from Russia to the other FSU countries were adversely affected by the Russian crisis. Specifically, trade credits and government loans from Russia to the other FSU countries halved during the Russian crisis, while Russian exporters tightened the terms of payments for delivery of goods and services supplied, in particular, energy products. Furthermore, Russia has been trying to reduce subsidies provided to some FSU countries in the form of cheap energy products. On the contrary, FDI inflows from other countries to the region increased during 2000–04 from 1993–97. A combination of these factors is likely to have weakened the growth linkages between Russia and other FSU countries.

Recent changes in the patterns of labor flows and remittances indicate that growth linkages between Russia and the other FSU countries may be strengthening again after the significant decline in the 1990s. The empirical analysis in this study suggests that transfers raise growth in the FSU countries. Therefore, it can be said that the recent increases in net transfers from Russia to other FSU countries have been stimulating growth in the recipient countries, perhaps through multiplier effects of consumption and increases in investments, including construction of new houses.

Appendix I. Reestimation of the Set of Equations in Shiells, Pani, and Jafarov (2005)

Shiells, Pani, and Jafarov (2005) (henceforth SPJ) specify an econometric model that includes standard growth determinants, as well as Russian economic growth, and allows for a possible shift in the regression coefficients following the Russian crisis.

Specifically, they estimate the following equation:

where γit is real GDP growth for country i in year t, γRt is real GDP growth for Russia in year t, xit is a vector of exogenous determinants of growth in country i, dt is a dummy variable equal to 0 prior to the Russian crisis and 1 thereafter, and µt and vit are disturbance terms.61

Shiells, Pani, and Jafarov (2005) run different specifications of the above equation to check for the robustness of their results. First, they try different exogenous determinants of growth (xit). Second, they repeat their estimations assuming break points in both 1998 and 1999 because of some uncertainty regarding the precise timing of the shift in output correlations.62 Third, they calculate Arellano-Bond estimators, given that the inclusion of a lagged dependent variable gives rise to a bias in standard estimators of either the fixed- or random effects model. However, Shiells, Pani, and Jafarov do not use lags of explanatory variables. Their results therefore should be interpreted with caution since lagged explanatory variables may be important.

Tables A1a and A1b present the results for the equations assuming that the structural break point was 1998.63 The specification presented in the first column includes lagged own-country growth, country dummies, the CPI, government expenditure in percent of GDP, the EBRD transition index, EU growth, the real exchange rate, Russian growth, the trade openness ratio, and interactions between a post-Russian crisis dummy variable and all of the other explanatory variables. The coefficient on Russian growth is quite substantial (1.08) and significant.64 The coefficient on the interaction of the Russian crisis dummy with Russian growth is -0.96 and highly significant. These results imply that, on average, a 1 percentage point increase in Russian growth was associated with a similar size increase in another country’s growth rate, holding other factors constant, before the Russian crisis. After the crisis, this effect dropped to 0.12 percentage points and was not significantly different from zero.

Table A1a.CIS and Baltic Countries: Coefficient Estimates in Real GDP Growth Regressions with Structural Break in 1998, 1993–2004 1/
(1)(2)(3)(4) 2/(5) 3/(6)(7)(8)(9)(10)
CONSTANT32.7312.6020.6947.6221.3926.0029.707.6433.817.86
(2.77)(2.09)(2.83)(3.87)(0.41)(2.10)(2.65)(1.28)(2.89)(0.52)
GR-10.110.390.390.23-0.530.120.230.440.110.04
(1.3)(5.173)(5.20)(2.50)(-1.71)(1.29)(2.57)(5.69)(1.31)(0.42)
CPI0.000.000.000.00-0.470.000.000.000.00
(-0.02)(3.41)(3.27)(0.41)(-2.23)(0.29)(0.08)(3.68)(-0.39)
INF-0.41
(-0.39)
EXP-0.05-0.19-0.160.100.00-0.03-0.12-0.11-0.04-0.12
(-0.44)(-2.69)(-1.95)(0.83)(0.01)(-0.24)(-1.10)(-1.60)(-0.37)(-1.03)
RI-3.932.64-0.66-8.97-3.66-2.52-2.033.64-4.56-3.37
(-1.05)(1.91)(-0.32)(-2.28)(-0.19)(-0.64)(-0.57)(2.62)(-1.19)(0.86)
GRRUS1.080.640.841.510.061.160.780.651.070.96
(5.65)(4.52)(4.94)(6.65)(0.15)(5.63)(3.91)(4.39)(5.56)(4.96)
EUGR-3.68-3.08-3.54-7.00-2.05-3.73-4.53-3.31-3.59
(-2.43)(-1.91)(-2.17)(-3.89)(-0.56)(-2.33)(-3.11)(-1.97)(-2.35)
WORLDGR3.89
(1.22)
RER-0.07-0.05-0.05-0.060.13-0.17-0.04-0.07-0.06
(-3.22)(-2.29)(-2.55)(-2.69)(0.70)(-4.82)(-2.01)(-3.20)(-2.66)
RRUS-0.010.14
(-0.50)(3.48)
OPEN-0.030.00-0.02-0.150.10-0.01-0.010.00-0.03-0.03
(-1.23)(0.02)(-1.03)(-0.56)(1.04)(-0.20)(-0.60)(-0.17)(-1.02)(-1.25)
IC22.15
(2.73)
D98-35.00-0.22-7.09-57.59-5.30-24.96-27.22-2.42-36.88-14.53
(-1.73)(0.03)(-0.74)(-2.37)(-0.09)(-1.16)(-1.39)(-0.34)(-1.83)(-0.70)
IC2×D98-0.55
(-0.504)
GR-1×D980.170.030.080.040.570.190.080.080.170.26
(1.08)(0.19)(0.52)(0.24)(1.36)(1.16)(0.51)(0.52)(1.09)(1.59)
CPI×D98-0.01-0.02-0.010.000.71-0.01-0.01-0.01-0.01
(-0.44)(-1.069)(-0.91)(0.25)(1.49)(-0.50)(-0.57)(-0.74)(-0.45)
INF×D98-0.38
(-011)
EXP×D98-0.26-0.01-0.06-0.38-0.19-0.24-0.140.02-0.27-0.19
(-1.42)(-0.10)(-0.48)(-1.85)(-0.48)(-1.16)(-0.74)(0.25)(-1.47)(-0.98)
RI×D988.19-4.27-0.3616.053.055.414.41-4.179.128.92
(1.23)(-2.34)(-0.14)(2.07)(0.15)(0.75)(0.67)(-2.34)(1.36)(1.37)
GRRUS×D98-0.96-0.41-0.62-1.37-0.11-1.05-0.65-0.41-0.95-0.86
(-4.42)(-2.38)(-3.13)(-5.30)(-0.23)(-4.49)(-2.80)(-2.27)(-4.36)(-3.69)
EUGR×D983.372.693.306.51-1.783.564.293.063.28
(2.06)(1.57)(1.90)(3.37)(0.47)(2.07)(2.73)(1.72)(1.97)
WORLDGR×D98-3.83
(-1.17)
RER×D980.080.020.020.09-0.330.220.010.080.07
(1.67)(.397)(0.423)(1.70)(-1.58)(3.27)(0.14)(1.67)(1.33)
RRUS×D98-0.01-0.19
(-0.18)(-3.20)
OPEN×D980.110.020.030.09-0.100.070.080.020.110.11
(2.15)(1.06)(1.33)(1.45)(-1.02)(1.27)(1.61)(0.95)(2.06)(2.12)
Source: Fund staff estimates.

Bold indicates statistically significant at the 1 percent level; italics indicates significance at the 5 percent level. t -statistics are in parentheses beneath the coefficient estimates.

Regression (4) includes the CIS countries only (i.e., the Baltics are excluded).

Regression (5) includes the Baltic countries only (i.e., the CIS countries are excluded).

Source: Fund staff estimates.

Bold indicates statistically significant at the 1 percent level; italics indicates significance at the 5 percent level. t -statistics are in parentheses beneath the coefficient estimates.

Regression (4) includes the CIS countries only (i.e., the Baltics are excluded).

Regression (5) includes the Baltic countries only (i.e., the CIS countries are excluded).

Table A1b.CIS and Baltic Countries: Coefficient Estimates in Real GDP Growth Regressions with Structural Break in 1998, 1993–2004 1/
(1)(2)(3)(4) 2/(5) 3/(6)(7)(8)(9)(10)
D2-12.19-10.88-10.85-9.20-12.46-12.10
D3-5.84-8.78-6.81-4.20-6.15-4.49
D42.291.806.772.100.99
D5-0.461.22-1.082.49-0.27-1.16
D6-11.23-6.97-7.39-7.37-11.03-12.56
D7-6.54-1.92-3.95-2.41-6.48-7.67
D8-1.97-0.251.63-2.06-3.23
D9-1.17-1.611.07-1.23-2.11
D10-15.82-14.66-15.44-10.80-16.25-16.06
D11-9.24-10.98-12.59-5.95-9.95-8.32
D12-15.59-15.30-15.17-12.45-15.80-15.36
D13-7.36-6.80-5.67-5.30-7.26-7.37
BALT6.42
CASIA-1.11
CAU0.27
D2×D9815.4616.4514.2112.5815.9015.14
D3×D9816.7428.3818.1815.4516.7615.06
D4×D98-8.52-8.07-13.18-8.60-6.15
D5×D98-3.85-6.63-3.25-6.81-4.15-3.12
D6×D989.505.125.705.669.2310.89
D7×D984.320.051.850.264.165.60
D8×D984.462.890.994.336.45
D9×D98-0.56-0.04-2.79-0.701.05
D10×D9813.4412.9313.208.4813.7213.96
D11×D983.7410.647.070.264.512.29
D12×D9819.1620.6318.9616.2119.3319.11
D13×D9811.5815.3910.129.7711.3811.22
BALT×D98-6.78
CASIA×D98-2.80
CAU×D98-3.05
Memorandum items:
GRRUS1.080.640.841.510.061.160.780.651.070.96
GRRUS×D98-0.96-0.41-0.62-1.37-0.11-1.05-0.65-0.41-0.95-0.86
Sum0.120.230.220.14-0.050.110.140.240.110.10
F-test1.284.9817.091.400.041.020.935.071.180.61
P-value0.260.030.000.240.840.310.340.030.280.44
Number of parameters42202436184244184242
Log likelihood-364.45-394.93-392.45-277.03-59.11-371.04-355.29-401.55-364.44-367.62
AIC2.862.972.992.881.972.962.773.032.862.91
R20.840.760.770.880.860.830.860.740.840.84
Adjusted R20.780.720.720.820.700.760.800.700.780.77
Source: Fund staff estimates.

Bold indicates statistically significant at the 1 percent level; italics indicates significance at the 5 percent level.

Regression (4) includes the CIS countries only (i.e., the Baltics are excluded).

Regression (5) includes the Baltic countries only (i.e., the CIS countries are excluded).

Source: Fund staff estimates.

Bold indicates statistically significant at the 1 percent level; italics indicates significance at the 5 percent level.

Regression (4) includes the CIS countries only (i.e., the Baltics are excluded).

Regression (5) includes the Baltic countries only (i.e., the CIS countries are excluded).

Based on the results in column (1), the coefficient on CPI inflation is not significant and is near zero, as is the coefficient on the interaction between CPI inflation and the Russian crisis dummy. While the coefficients on government expenditure (as a percent of GDP) and its interaction with the Russian crisis dummy have negative signs, they are statistically insignificant. Growth in the EU has a negative coefficient (−3.7) and is significant (at the 5 percent level) prior to 1998, while in the 1998–2004 period it changes its sign (3.7).

Estimation of an alternative specification, including initial condition measure IC2 but not country effects, is shown in column (2) of Tables A1a and A1b. The results are broadly similar to those obtained above. The estimated coefficient on Russian growth is 0.64 and highly significant prior to 1998, while thereafter it fell to 0.23. However, in contrast with the original paper, the variable remains significant at the 5 percent level. Estimated coefficients corresponding to the CPI and government expenditure in percent of GDP are statistically significant in this specification, although the coefficient on the CPI is still near zero.

A regression equation including regional dummy variables (Baltics; the Caucasus and Moldova; and Central Asia—Belarus and Ukraine constitute the reference group) contains the same variables as in column (1) except that all terms involving country dummies are omitted, and the following variables involving regional dummies are added: (i) the regional dummies themselves; and (ii) the interactions of regional dummies with the Russian crisis dummy. Results reported in column (3) of Table A1 are broadly similar to the results in column (2).

The results are insensitive to the choice of whether the Baltics are included in the sample, which suggests importantly that the paper’s findings are quite robust with respect to changes in the country sample. Column (4) of Table A1 presents estimates based on the previous specification but excluding the Baltics. The coefficient on growth in Russia is 1.51 and highly significant, while the coefficient on the interaction of the Russian crisis dummy with Russian growth is -1.37 and highly significant; their difference is 0.14 and insignificant. For completeness, column (5) includes estimates using data only for the Baltics. In this case, the coefficients on growth in Russia and its interaction with the Russian crisis dummy are both insignificant.

Two other specifications comprise a CPI-based bilateral real exchange rate vis-à-vis Russia either instead of, or in addition to, the multilateral real effective exchange rate index used in the previous specifications. Results from these regressions—reported in columns (6) and (7) of Tables A1a and A1b—are very similar to those reported above for the key variables of interest. In column (7), for instance, the coefficient on Russian growth is 0.78 and highly significant prior to 1998, while it falls to 0.14 and becomes insignificant thereafter.

Tables A1a and A1b include estimates for a variety of other specifications, results of which are all broadly similar for the key variables of interest—Russian growth and its interaction with the crisis dummy. Column (8) provides estimates for a specification that excludes both country fixed effects and initial conditions. Possible nonlinearity in the response of growth to inflation is explored in column (9) by including INF, the natural logarithm of percent changes in CPI inflation, in place of CPI. The coefficient on (log) CPI inflation is still insignificant, while the coefficients on Russian growth are similar to those reported earlier, indicating that the results are robust to changes in the functional form. Column (10) includes estimates based on substituting world growth for EU growth. While the coefficient on world growth is insignificant, once again the coefficients on Russian growth are similar to the earlier results.

Tables A2a and A2b present results for the same specifications as in Tables A1a and A1b but assuming that the break occurred in 1999 rather than 1998. The fit of these regressions, with the exceptions of the second, fifth (the Baltics-only regression), and eighth specifications, are worse than those of regressions using a break point of 1998, as reflected in the lower values of the log-likelihood function, the higher values of the AIC, and the smaller number of significant t-statistics. While Russian growth and its interaction with the Russian financial crisis dummy are not significant in the fifth specification, assuming a break point of 1999, in the case of the second and eighth specifications, estimations using a break point of 1998 yield a larger number of significant variables.

Table A2a.Coefficient Estimates in Real GDP Growth Regressions with Structural Break in 1999, CIS and Baltic Countries, 1993–2004 1/
(1)(2)(3)(4) 2/(5) 3/(6)(7)(8)(9)(10)
CONSTANT16.8016.6921.2022.8949.208.0630.0211.3120.1923.36
(1.86)(3.12)(3.42)(2.27)(1.26)(0.88)(3.30)(2.10)(2.13)(2.00)
GR-10.260.440.480.36-0.710.280.260.490.220.20
(3.65)(6.95)(7.35)(4.28)(-3.53)(3.77)(3.91)(7.51)(3.09)(2.79)
CPI0.000.000.000.00-0.550.000.000.000.00
(1.70)(3.46)(3.51)(2.17)(-3.27)(2.04)(2.12)(3.73)(1.40)
INF-0.33
(-0.32)
EXP-0.12-0.23-0.24-0.050.19-0.14-0.18-0.12-0.09-0.14
(-1.14)(-3.10)(-3.29)(-0.38)(0.72)(-1.18)(-1.78)(-2.03)(-0.77)(-1.28)
RI1.871.300.040.43-13.724.24-1.332.37-0.100.31
(0.66)(1.10)(0.29)(0.14)(-0.94)(1.47)(-0.48)(1.98)(-0.03)(0.10)
GRRUS0.780.690.711.000.290.800.720.700.780.93
(5.37)(5.19)(4.97)(5.57)(1.35)(5.20)(5.24)(5.01)(5.06)(5.19)
EUGR-1.61-2.60-2.39-3.52-4.35-1.40-3.81-3.07-1.29
(-1.16)(-1.80)(-1.61)(-2.09)(-1.83)(-0.93)(-2.71)(-2.04)(-0.92)
WORLDGR-1.50
(-1.23)
RER-0.07-0.05-0.06-0.070.24-0.16-0.05-0.08-0.74
(-1.44)(-2.78)(-2.88)(-3.08)(1.74)(-5.41)(-2.30)(-3.59)(-3.46)
RRUS0.000.13
(-0.09)(3.97)
OPEN-0.040.00-0.01-0.030.13-0.02-0.040.00-0.04-0.05
(-1.84)(0.06)(-0.69)(-1.36)(1.74)(-0.96)(-1.69)(-0.11)(-1.65)(-2.09)
IC22.53
(3.61)
D99-39.58-11.47-17.05-52.66-35.94-28.11-49.60-10.54-42.96-48.70
(-1.77)(-1.44)(-1.83)(-1.52)(-0.78)(-1.19)(-2.32)(-1.53)(-1.91)(-2.16)
IC2×D99-1.64
(-1.49)
GR-1 ×D990.040.020.04-0.030.870.040.040.040.080.06
(0.22)(0.13)(0.22)(-0.13)(2.58)(0.21)(0.21)(0.24)(0.41)(0.33)
CPI×D990.00-0.01-0.010.000.83-0.010.00-0.010.00
(-0.25)(-0.83)(-0.75)(-0.11)(1.83)(-0.31)(-0.30)(-0.70)(-0.26)
INF×D99-0.76
(-0.21)
EXP×D99-0.050.110.14-0.09-0.270.050.130.07-0.08-0.06
(-0.21)(1.02)(1.09)(-0.32)(-0.81)(0.20)(0.55)(0.73)(-0.38)(-0.27)
RI×998.06-2.00-0.5411.8011.644.239.58-2.4210.0211.11
(1.08)(-1.16)(-0.28)(1.21)(0.71)(0.53)(1.33)(-1.45)(1.32)(1.54)
GRRUSCPI×D99-0.240.130.12-0.480.47-0.28-0.240.13-0.24-0.93
(-0.56)(0.31)(0.29)(-0.96)(0.81)(-0.63)(-0.60)(0.30)(-0.55)(-5.19)
EUGRCPI×D990.971.451.332.862.800.863.232.040.69
(0.58)(0.87)(0.78)(1.43)(1.02)(0.48)(1.96)(1.17)(0.41)
WORLDGRCPI×D991.75
(1.32)
RERCPI×D990.060.000.000.06-0.440.20-0.020.060.07
(1.06)(-0.02)(-0.09)(1.06)(-2.72)(2.73)(-0.35)(1.10)(1.21)
RRUSCPI×D99-0.03-0.18
(-0.73)(-3.10)
OPENCPI×D990.080.020.020.07-0.130.050.070.020.080.10
(1.56)(0.82)(1.09)(1.01)(-1.58)(0.88)(1.41)(0.78)(1.49)(1.88)
Source: Fund staff estimates.

Bold indicates statistically significant at the 1 percent level; italics indicates significance at the 5 percent level. t-statistics are in parentheses beneath the coefficient estimates.

Regression (4) includes the CIS countries only (i.e., the Baltics are excluded).

Regression (5) includes the Baltic countries only (i.e., the CIS countries are excluded).

Source: Fund staff estimates.

Bold indicates statistically significant at the 1 percent level; italics indicates significance at the 5 percent level. t-statistics are in parentheses beneath the coefficient estimates.

Regression (4) includes the CIS countries only (i.e., the Baltics are excluded).

Regression (5) includes the Baltic countries only (i.e., the CIS countries are excluded).

Table A2b.Coefficient Estimates in Real GDP Growth Regressions with Structural Break in 1999, CIS and Baltic Countries, 1993–2004 1/
(1)(2)(3)(4) 2/(5) 3/(6)(7)(8)(9)(10)
D2-4.40-2.96-4.26-4.36-6.54-5.26
D30.91-0.420.780.87-1.970.17
D4-2.40-0.62-1.00-2.05-1.05
D5-4.38-3.63-3.86-4.02-2.60-4.49
D6-10.12-8.21-8.30-8.76-10.04-10.17
D7-6.09-4.56-4.78-4.85-6.06-5.86
D8-2.94-1.63-1.91-3.00-2.14
D9-2.64-2.22-2.10-2.44-1.95
D10-12.13-11.82-10.91-11.20-14.31-12.09
D11-5.43-4.81-5.05-5.21-8.59-6.18
D12-9.50-8.96-8.79-8.99-11.89-9.99
D13-3.54-3.54-3.34-3.35-4.51-3.83
BALT3.14
CASIA-3.15
CAU-1.92
D2×D9910.2711.149.549.4512.439.72
D3×D9919.2425.3716.8314.6621.6612.34
D4×D99-10.51-12.33-12.95-10.57-9.95
D5×D99-0.43-1.30-0.60-0.37-2.330.13
D6×D999.697.467.948.249.579.91
D7×D995.022.833.202.474.913.63
D8×D99-5.77-3.33-4.11-0.38-2.14
D9×D99-5.40-6.30-7.34-5.48-5.62
D1 0CPI×D9910.999.149.408.6613.089.75
D11CPI×D998.2011.479.269.2311.218.94
D12CPI×D9914.4213.4712.4211.5216.7812.28
D13CPI×D9912.7418.1311.2610.6413.318.57
BALTCPI×D99-2.77
CASIACPI×D991.11
CAUCPI×D990.84
Memorandum items:
GRRUS0.780.690.711.000.290.800.720.700.780.93
GRRUSCPI×D99-0.240.130.12-0.480.47-0.28-0.240.13-0.24-0.93
Sum0.540.820.830.520.750.520.480.830.540.00
F -test1.854.404.331.201.991.541.594.131.752.18
P-value0.180.040.040.280.180.220.210.040.190.14
Number of parameters42202436184244184242
Log likelihood-368.98-392.25-392.55-285.78-55.62-376.36-357.66-400.03-370.90-368.58
AIC2.932.932.993.041.763.032.803.012.952.92
R20.830.770.770.860.890.810.860.740.830.83
Adjusted R20.760.730.720.790.760.740.790.710.760.77
Source: Fund staff estimates.

Bold indicates statistically significant at the 1 percent level; italics indicates significance at the 5 percent level.

Regression (4) includes the CIS countries only (i.e., the Baltics are excluded).

Regression (5) includes the Baltic countries only (i.e., the CIS countries are excluded).

Source: Fund staff estimates.

Bold indicates statistically significant at the 1 percent level; italics indicates significance at the 5 percent level.

Regression (4) includes the CIS countries only (i.e., the Baltics are excluded).

Regression (5) includes the Baltic countries only (i.e., the CIS countries are excluded).

While the coefficients are generally less precisely estimated when using a break point of 1999, the results are very similar in many respects to those presented in Tables A1a and A1b above. In particular, the coefficient on Russian growth is broadly similar in magnitude to the estimates based on a 1998 break point and highly significant in all but one specification. The coefficient on the interaction of the Russian crisis dummy with Russian growth ranges widely and is significant in only one of the specifications. Taken together, these results provide support for the choice of 1998 as the structural break point.

Since the inclusion of a lagged dependent variable in the error components model generates a bias in the LSDV estimators, the results based on LSDV estimation, presented in column (1), are compared with the results based on Arellano-Bond estimation in columns (11) and (12) in Table A3.65 As can be seen from this table, coefficients on Russian growth are broadly comparable to those obtained using the LSDV estimation. Arellano-Bond estimates of the coefficient on the interaction between the crisis dummy and Russian growth are also comparable to the LSDV estimates and are significant in both specifications.66 Results of the Sargan test do not reject the null hypothesis that the overidentifying restrictions underlying the Arellano and Bond (1991) estimation method are satisfied in the twelfth specification, suggesting that the instruments are valid. Finally, the null hypotheses of second-order serially uncorrelated errors are not rejected, fulfilling a necessary condition for consistency of the Arellano-Bond estimation procedure. Column (13) presents Arellano-Bond estimates that also correct for possible endogeneity of the explanatory variables CPI and EXP, using one-period lagged values of these variables as instruments. These estimates are very similar to estimates based on the assumption that the explanatory variables are exogenous.

Table A3.Arellano—Bond Estimates of Real GDP Growth Regressions with a Structural Break in 1998, CIS and Baltic Countries, 1993-2004 1/
LSDV EstimatesStrictly Exogenous Explanatory VariablesEndogeneity Correction 2/
(1)(11)(12)(13)
Coefficientt-statisticCoefficientt-statisticCoefficientt-statisticCoefficientt-statistic
CONSTANT32.732.770.471.590.351.420.271.05
GR -10.111.300.314.420.456.640.446.57
CPI0.00-0.020.001.040.002.430.002.93
EXP-0.05-0.44-0.07-0.72-0.21-2.17-0.31-2.81
RI-3.93-1.051.100.38
GRRUS1.085.650.684.430.674.740.755.60
EURGR-3.68-2.43-3.10-2.38
WORLDGR-0.02-0.030.050.08
RER-0.07-3.22-0.05-2.30
OPEN-0.03-1.23-0.01-0.39
D98-35.00-1.736.810.98-3.25-2.30-3.11-2.24
GR -1 ×D980.171.08-0.20-1.40
CPICPI×D98-0.01-0.44-0.02-1.23
EXPCPI×D98-0.26-1.42-0.24-2.20
RICPI×D988.191.23-5.51-2.76
GRRUSCPI×D98-0.96-4.42-0.63-3.87-0.52-3.06-0.61-3.78
EURGRCPI×D983.372.062.842.07
RERCPI×D980.081.670.030.74
OPENCPI×D980.112.150.082.77
Memorandum items:
Sargan test of overidentifying restrictions83.660.01 3/40.200.92 3/96.311.00 3/
Arellano-Bond test for AR(1)-1.920.05 3/-3.370.00 3/-3.270.00 3/
Arellano-Bond test for AR(2)-0.640.52 3/0.350.72 3/0.250.80 3/
Source: Fund staff estimates.

Bold indicates statistically significant at the 1 percent level; italics indicates significance at the 5 percent level.

Coefficient estimates in this column assume that CPI and EXP are endogenous. First-order lagged values of these variables are used as instruments.

These figures refer to p-values instead of t -statistics.

Source: Fund staff estimates.

Bold indicates statistically significant at the 1 percent level; italics indicates significance at the 5 percent level.

Coefficient estimates in this column assume that CPI and EXP are endogenous. First-order lagged values of these variables are used as instruments.

These figures refer to p-values instead of t -statistics.

References

    AddletonJonathan S.1992Undermining the Centre: The Gulf Migration and Pakistan.(Oxford:Oxford University Press.

    AdelmanIrma J. EdwardTaylor and StephenVogel1988“Life in a Mexican Village: A SAM Perspective,”Journal of Development StudiesVol. 25pp. 5–24.

    • Search Google Scholar
    • Export Citation

    AgénorPierre-Richard1996“The Labor Market and Economic Adjustment,”Staff PapersInternational Monetary Fund,Vol. 43(March)pp. 261–335.

    • Search Google Scholar
    • Export Citation

    AgénorPierre-RichardC. JohnMcDermott and Eswar S.Prasad1999“Macroeconomic Fluctuations in Developing Countries: Some Stylized Facts,”IMF Working Paper 99/35(Washington:International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    AndrienkoYuri and SergeiGuriev2004“Determinants of Interregional Mobility in Russia,”Economics of TransitionVol. 12pp. 1–27.

    AndrienkoYuri and SergeiGuriev2005“Understanding Migration in Russia,”World Bank Policy Note.Available via Internet: http://www.cefir.ru/download.php?id=216.

    • Search Google Scholar
    • Export Citation

    ArellanoManuel and StephenBond1991“Some Tests of Specification for Panel Data: Monte Carlo Evidence and an Application to Employment Equations,”Review of Economic StudiesVol. 58pp. 277–97.

    • Search Google Scholar
    • Export Citation

    AroraVivek and AthanasiosVamvakidis2005“How Much Do Trading Partners Matter for Economic Growth?”Staff PapersInternational Monetary Fund,Vol. 52(No. 1)pp. 24–40.

    • Search Google Scholar
    • Export Citation

    BackusDavid K. and Patrick J.Kehoe1992“International Evidence on the Historical Properties of Business Cycles,”American Economic ReviewVol. 82pp. 864–88.

    • Search Google Scholar
    • Export Citation

    BackusDavid K. and Patrick J.Kehoe and Finn E.Kydland1992“International Real Business Cycles,”Journal of Political EconomyVol. 100pp. 745–75.

    • Search Google Scholar
    • Export Citation

    BaldwinR. and E.Seghezza1996“Testing for Trade Induced-Investment-Led Growth,”NBER Working Paper No. 5416(Cambridge, Massachusetts:National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation

    BerengautJulian ErikDe Vrijer KatrinElborgh-Woytek MarkLewis and BogdanLissovolik2002“An Interim Assessment of Ukrainian Output Developments, 2000–01,”IMF Working Paper 02/97(Washington:International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    BernardB. Andrew and J. BradfordJensen2004“Entry, Expansion, and Intensity in the US Export Boom, 1987–1992,”Review of International EconomicsVol. 12(No. 4)pp. 662–75.

    • Search Google Scholar
    • Export Citation

    BourdetYves and HansFalck2003“Emigrants, Remittances, and Dutch Disease in Cape Verde,”Working Paper Series 2003: 11(Kristianstad, Sweden:Kristianstad University College;available via Internet: http://hkep1fs.hkr.se/upload/E/Working%20papers/wp2003_11.pdf)

    • Search Google Scholar
    • Export Citation

    ChamiRalph ConnelFullenkamp and SamirJahjah2003“Are Immigrant Remittances Flows a Source of Capital for Development?,”IMF Working Paper 03/189(Washington:International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    CamposNauro F. and YukoKinoshita2003“Why Does FDI Go Where it Goes? New Evidence from the Transition Economies,”IMF Working Paper 03/228(Washington:International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    Central Bank of Russia2005“Cross-Border Remittances: Russian Experience,”mimeo.

    CucMilan ErikLundback and EdgardoRuggiero2005Migration and Remittances in Moldova(Washington:International Monetary Fund).

    de MolloLuiz R.1999“Foreign Direct Investment-led Growth: Evidence from Time Series and Panel Data,”Oxford Economic PapersVol. 52(January)pp. 133–51.

    • Search Google Scholar
    • Export Citation

    DesaiMihir DeveshKapur and JohnMcHale2001“The Fiscal Impact of the Brain Drain: Indian Emigration to the U.S.,”prepared for the Third Annual NBER-NCAER conference,Neemrana, India(available via Internet: http://www.wcfia.harvard.edu/seminars/pegroup/Desai-Kapur.pdf).

    • Search Google Scholar
    • Export Citation

    DollarDavid1992“Outward-Oriented Developing Countries Really Do Grow More Rapidly: Evidence from 95 LDCs, 1976–1985,”Economic Development and Cultural Change(April)pp. 523–44.

    • Search Google Scholar
    • Export Citation

    DoyleBrian M. and JonFaust2002“An Investigation of Co-Movements Among the Growth Rates of the G-7 Countries,”Federal Reserve Bulletin(October)pp. 427–37.

    • Search Google Scholar
    • Export Citation

    EdwardsSebestian1992“Trade Orientation, Distortions, and Growth in Developing Countries,”Journal of Development EconomicsVol. 39(July)pp. 31–57.

    • Search Google Scholar
    • Export Citation

    European Bank for Reconstruction and Development2000“Human Dimensions of Transition,”Chapter 1 in Transition Report 2000(London:European Bank for Reconstruction and Development).

    • Search Google Scholar
    • Export Citation

    European Bank for Reconstruction and DevelopmentTransition Report various issues(London:European Bank for Reconstruction and Development).

    • Search Google Scholar
    • Export Citation

    EU Commission2004“Russia - WTO: EU-Russia deal brings Russia a step closer to WTO membership.”Available via Internet: http://europa.eu.int/comm/external_relations/russia/intro/ip04_673.htm.

    • Search Google Scholar
    • Export Citation

    FilerRandall K. ThorvaldurGylfason StepanJurajda and JanetMitchell2000“Markets and Growth in the Post-Communist World.”Available via Internet: http://www.gdnet.org/pdf/445_Glyfason-Mitchel-Filer.pdf.

    • Search Google Scholar
    • Export Citation

    FriebelGuido and SergeiGuriev2005“Attaching Workers Through In-Kind Payments: Theory and Evidence from Russia,”World Bank Economic ReviewVol. 19(No. 2)pp. 175–202.

    • Search Google Scholar
    • Export Citation

    Global Commission on International Migration (GCIM)2005“Migration in an Interconnected World: New Directions for Action,”Report of the Global Commission on International Migration.Available via Internet: http://www.gcim.org/en/finalreport.html.

    • Search Google Scholar
    • Export Citation

    GrossmanGene M. and ElhananHelpman1991Innovation and Growth in the Global Economy(Cambridge, Massachusetts and London:MIT Press).

    GurievSergei and Barry W.Ickes1999“Barter in Russian Enterprises: Myths vs. Empirical Evidence,”Russian Economic TrendsVol 8(No. 2)pp. 6–13.

    • Search Google Scholar
    • Export Citation

    HavrylyshynOleh and Lucio VinhasDe Souzaforthcoming,Growth Resumption in the CIS Countries(Washington:International Monetary Fund).

    HavrylyshynOleh and Lucio VinhasDe Souza and Ronvan Rooden2000“Institutions Matter in Transition, but So Do Policies,”IMF Working Paper 00/70(Washington:International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    HelblingThomas and TamimBayoumi2003“Are They All in the Same Boat? The 2000–2001 Growth Slowdown and the G−7 Business Cycle Linkages,”IMF Working Paper 03/46(Washington:International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    IETRussian Bulletin of Business Cycle Surveys(in Russian), various issues(Moscow, Russia:Institute for the Economy in Transition).

    International Monetary Fund2005“Two Current Issues Facing Developing Countries,” inWorld Economic OutlookApril 2005Chapter II,(Washington:International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    IvakhnyukIrina2003“Dve Migratsionnye Sistemy v Evrope: Tendentsii Razvitiya i Perspektivy Vzaimodeystviya,”Available via Internet: http://www.archipelag.ru/agenda/povestka/povestka-immigration/strategii/dve_sistemi/.

    • Search Google Scholar
    • Export Citation

    JafarovEtibar and KatsiarynaSvirydzenkaforthcoming,“Patterns of Trade of the FSU Countries,”mimeo.

    KireyevAlexei2006“The Macroeconomics of Remittances: The Case of Tajikistan,IMF Working Paper 06/2(Washington:International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    KoseM. Ayhan Eswar E.Prasad and Marco E.Terrones2003“How Does Globalization Affect the Synchronization of Business Cycles?”IMF Working Paper 03/27(Washington:International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    KwonGoohoon and AntonioSpilimbergo2005“Russia’s Regions: Income Volatility, Labor Mobility, and Fiscal Policy,”IMF Working Paper No. 05/185(Washington:International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    LevineRoss and DavidRenelt1992“A Sensitivity Analysis of Cross-Country Growth Regressions,”American Economic ReviewVol. 82(September)pp. 942–963.

    • Search Google Scholar
    • Export Citation

    MaddalaG.S. and In-MooKimUnit Roots Cointegration and Structural Change(Cambridge:Cambridge University Press).

    MeyerJean Baptiste and MercyBrown1999“Scientific Diasporas: A New Approach to the Brain Drain,”UNESCO Discussion Paper No. 41Management of Social Transformations(Paris).

    • Search Google Scholar
    • Export Citation

    OwenDavid and David O.Robinson2003Russia Rebounds(Washington:International Monetary Fund).

    OomesNienke and OksanaDynnikova2006“The Utilization-Adjusted Output Gap: Is the Russian Economy Overheating?”IMF Working Paper 06/68(Washington:International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    RobertsMark and James R.Tybout1997“The Decision to Export in Colombia: an Empirical Model of Entry with Sunk Costs,”American Economic ReviewVol. 87(September)pp. 545–64.

    • Search Google Scholar
    • Export Citation

    RodriguezFrancisco and DaniRodrik2001“Trade Policy and Economic Growth: A Skeptic’s Guide to the Cross-National Evidence,” inMacroeconomics Annual2000,ed. byBenBernanke and KennethRogoff(Cambridge, Massachusetts:MIT Press for National Buearau of Economic Research).

    • Search Google Scholar
    • Export Citation

    Russian Economic Barometer2004“Outline of REB Surveys,”Russian Economic BarometerVol. 13(No. 4)pp. 38–40.

    SachsJeffrey D. and AndrewWarner1995“Economic Reform and the Process of Global Integration,”Brookings Papers on Economic ActivityBrookings Instituion,pp. 1–118.

    • Search Google Scholar
    • Export Citation

    ShiellsClinton R. MarcoPani and EtibarJafarov2005“Is Russia Still Driving Regional Economic Growth?”IMF Working Paper 05/192(Washington:International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    World Bank2005aDying Too Young; Addressing Premature Mortality and Ill Health Due to Non-Communicable Diseases and Injuries in the Russian Federation.

    • Search Google Scholar
    • Export Citation

    World Bank2005bThe Labor Market in Russia: Moving from Crisis to Recoverya copublication of Izdatelstvo Ves Mir and the World Bank.

32Prepared by Etibar Jafarov.
33Researchers studying the international correlation of output changes have mainly analyzed the transmission of business cycles among the industrial countries, although some papers have studied the business cycles of developing countries (see, for instance, Backus, Kehoe, and Kydland, 1992; Backus and Kehoe, 1992; Agénor, McDermott, and Prasad, 1999; Doyle and Faust, 2002; and Helbling and Bayoumi, 2003). Given difficulties in defining business cycles in transition economies, SPJ, as well as this paper, focus on growth correlations.
34Turkmenistan is excluded from the analysis because of data availability problems.
35See Grossman and Helpman (1991) and Baldwin and Seghezza (1996).
36This process was mainly related to (i) significant inefficiencies in production and trade within the USSR; (ii) the collapse of the ruble zone and introduction of new currencies; (iii) the imposition of trade barriers and political problems among some FSU countries; and (iv) increased access to foreign markets (Jafarov and Svirydzenka, forthcoming).
37All the FSU countries, except Estonia, Latvia, and Lithuania, are members of the CIS. In 2005, Turkmenistan ceased its permanent membership and became an associated member of this organization.
38From August 1998 to August 1999, the Russian ruble lost about 75 percent of its value against the U.S. dollar. Since many FSU countries were de facto targeting the U.S. dollar (or other “hard currencies”), their currencies initially appreciated against the Russian ruble. Later, most CIS countries allowed their currencies to depreciate against the U.S. dollar.
39This shift was facilitated by substantial reductions in public expenditures.
40In 7 out of 13 countries, the ratios of exports to GDP increased during the 2000–04 period from those in the precrisis period (Table 2).
41During 2000–04, the industrial sector shed labor in all the FSU countries from the levels of the 1993–97 period, while the share of the agriculture sector in total employment increased in Azerbaijan, Armenia, Kazakhstan, the Kyrgyz Republic, Moldova, and Ukraine. The share of the service sector increased in Azerbaijan, Armenia, Belarus, Baltic countries, and Ukraine. Also, in 6 out of the 13 countries, the increases in the share of agriculture in total employment coincided with declines in the share of this sector in GDP, while declines in the share of the industrial sector in total employment coincided with increases in the share of this sector in GDP.
42Guriev and Ickes (1999) argue that barter reduces transparency in governing enterprises and the economy and is associated with a lack of restructuring.
43Roberts and Tybout (1997) and Bernard and Jensen (2004) suggest that sunk entry costs affect trade patterns.
44Caution is needed in interpreting changes in the real effective exchange rate (REER) indices. First, these indices are based on official exchange rates, whereas some FSU countries (e.g., Belarus and Uzbekistan) have operated under systems of multiple exchange rates. Second, these indices are based on trade weights at a certain time and do not reflect changes in the trade structure of these countries. Third, while the consumer price index (CPI) may be a highly inaccurate index of price competitiveness, possibly more appropriate indicators, such as the producer price index (PPI), are not available for all countries and all years in the sample.
45On financial flows from Russia to the other FSU countries, see next section.
46Prices for gas have been raised also for domestic use in Russia, but much less than the increases for the other FSU countries. Under an EU-Russia agreement signed in May 2004, Russia agreed to gradually increase gas prices for industrial Russian users from US$27–28 in 2004 to between US$49–57 by 2010 (EU, 2004). It is not clear if these prices will be revised due to a significant increase in market prices since then. For comparison, starting in 2006, Russia charges Georgia at US$110 and Moldova at US$160 per thousand cubic meters of its gas exports.
47The contribution of exports to Russia from one of the CIS and Baltic countries to growth in the latter is estimated as follows: 100× Δ [(XRt/Xt)xt]/yt−1where XRt and Xt are the values of exports to Russia and total exports of merchandise, respectively, xt is real total exports of goods and services, and γt−1 is real GDP.
48Strictly, only net exports should be compared with GDP. Caution is needed in comparing exports and GDP because the former includes the imported intermediate inputs used to produce exports, whereas the latter includes only value added. In addition, the data underlying these decompositions have substantial shortcomings, such as weaknesses in the expenditure decomposition of GDP. Furthermore, the estimates are based on the assumption that deflators for exports to Russia and to the rest of the world are same.
49Gravity models relate trade between countries to income of the countries and distances between them. Jafarov and Svirydzenka (forthcoming) also control for a number of other variables that are believed to affect trade flows.
50Empirical studies suggest that the main determinants of FDI in transition economies are institutions, natural resources, trade openness, market size, agglomeration economies, and labor costs. In the case of the CIS countries, abundant natural resources and economic reforms are the main determinants of FDI inflows (Campos and Kinoshita, 2003).
51Total foreign financing is defined as the difference between the current account balance and international reserve accumulation.
52Anecdotal evidence suggests that a significant share of FDI resulted from the repatriation of capital that had fled from these countries.
53Labor market segmentation, impediments to labor mobility, and other rigidities hinder growth (Agenor 1996; and Filer and others, 2000).
54This number may include double counting since not all outflows are to non-FSU countries.
55Among the CIS countries, only Belarus registered net immigration from Russia. Anecdotal evidence suggests that this outcome was mainly due to the return of a large number of military servicemen of Belarusian nationality upon their retirement from service in the Russian army.
56Cuc, Lundback, and Ruggiero (2005) estimate remittances, including workers’ compensation, to Moldova at 27 percent of GDP in 2004.
57The experience of other countries suggests that people who stay abroad for short periods tend to remit more than those who stay longer, since the latter establish bonds in their host countries, and have more options to invest in their host countries, including in real estate (GCIM, 2005).
58The remittances-to-GDP ratio was not included because of data problems. Specifically, a large number of observations are missing for this variable, and data do not reflect actual flows for most of the 1990s (see also the penultimate paragraph in Section D).
59Results obtained by including real effective exchange rates should be interpreted with care, owing to several shortcomings in the calculation of these indices. See footnote 13.
60The fourth and fifth specifications in SPJ, which use subsets of the data set, and which are similar to the third specification in that paper, are not reestimated here.
61Under the assumption of fixed country effects, the restrictions that the sums of µi s and dtµ s are equal to zero need to be imposed on the estimation of equation (A1) to avoid the dummy variable trap. Initial conditions must also be excluded from xit under the assumption of fixed effects (but can be included under the assumption of random effects) since the initial conditions vary across countries but not over time and hence are perfectly collinear with the country effects.
62Maddala and Kim (1998, p. 398) argue that prior information on the regime switch point should be used if it is available—thereby raising the question of whether there was a structural change around that period—rather than simply endogenizing the break point.
63The regressions in Tables A1 and A2 use LSDV estimators while Table 11 of the main text uses GLS estimators (see the second paragraph of Section E of the main text of this paper).
64This result is broadly consistent with the finding in Arora and Vamvakidis (2005) that a 1 percentage point increase in economic growth of trading partners is correlated with as much as a 0.8 percentage point increase in domestic growth.
65The Arellano and Bond (1991) estimates presented in Table A3 correspond to a one-step procedure, using one-period lags of the independent variables as instruments.
66Because the Arellano-Bond procedure uses first differences of strictly exogenous regressors as instruments, time-invariant strictly exogenous regressors, such as the country fixed effects, drop out.

Other Resources Citing This Publication