Hence, whilst proceeding to the empirical analysis of convergence processes between Russia’s regions, it becomes possible to conclude that at present there exist a few convergence concepts in respect to different countries or regions of a given country, which consequently implies a few methodologies of the analysis of convergence. First, the simplest way to analyze the presence of a convergence in the economic development rates implies the perspective of reduction in an inequality in terms of per capita GDP (GRP) between a group of regions (i.e. to consider the –convergence concept). This particular approach implies most frequently an analysis of various indicators that characterize the level of inequity of the given countries or regions in terms of some index of income, most frequently – GDP (GRP). Secondly, to test the - convergence concept, researchers most frequently employ Quah (1993, 1995, 1996).

cross–section analysis across a given sample of countries or regions. Thirdly, testing of convergence concepts on the basis of the analysis of time series of income indices and studying their dynamic characteristics appears somewhat more sophisticated from the methodological perspective. However, in our case the absence of relatively long series of data across Russia’s regions does not allow to employ the time series method to test a convergence hypothesis.

2.2. An Empirical Testing of Convergence Concepts in Respect to the RF Regions 2.2.1. Comparative Analysis of the RF Regions by GRP over the Period between 1994 through For the purpose of statistical analysis we used statistical data of the RF Goskomstat across 88 regions21 of the Russian Federation over the period between 1994 through 2002. Before an empirical testing of convergence concepts in respect to Russia’s regions, we have conducted a comparative analysis of regions by the size of their regional income. For this purpose, we employed per capita GRP in the 1994 prices. The adjustment of GRP per capita values to the 1994 prices was made by means of regional CPI=s. Bar charts of dispersion of regions by GRP per capita values between 1994 to 2002 are given in Fig. 2.1, while main statistical characteristics for each years are presented in Table 2.1.

Because of the absence of the data on a number of its macroeconomic indicators, Chechen Republic was excluded from the group of examined regions.

0 2000 4000 6000 6000 8000 10000 12000 14000 16000 18000 13000 20000 22000 24000 26000 28000 30000 6000 16000 22000 24000 28000 32000 2000 50 0 Fig 2.1. Dispersion of Russia’s Regions by level of GRP Per Capita in Constant Prices (as Rb. Th.) in 1994–Table 2.Main Statistical Characteristics of GRP Series between 1994 to Minimum value Maximum value Average value Median Standard Bias 1994 458.12 30802.78 4538.58 2867.05 5694.1995 449.53 26318.35 4124.38 3119.03 3791.1996 414.31 30096.50 4567.54 3293.81 4508.1997 414.33 30339.85 4610.18 3281.92 4534.1998 734.23 19809.48 3018.30 2159.68 3050.1999 641.83 37185.51 3820.46 2440.29 5000.2000 1133.73 36972.71 4459.18 3007.17 5666.2001 948.21 39073.32 4764.78 3279.82 5997.2002 622.81 63690.17 5831.56 3590.75 9171. The above bar charts allows to note that throughout the whole period in question, roughly as many as half of the regions had their GRP per capita at the level not more than Rb. 2,000, while some two–thirds of regions – at the level of not more than 4,000. Given that, the dispersion of regions in terms of GRP per capita level for each of the years in question appears unimodal, which, following Quah (see above) testifies to the presence of convergence processes.

Overall, during the whole period in question the average GRP value was at the level Rb. 4,500 per capita, with a notable fall in this index down to 3,000 in 1998, in the wake of the 1998 crisis.

However, since 1999 the regional per capita incomes have been growing and by 2002 they practically doubled vis--vis the minimum value and reached Rb. 5,800. Notably, as far as the time period starting from 1999 is concerned, the rise in the average GRP value per capita was also accompanied by a steady rise in the median GRP per capita value. In other words, the rise in the average living standards was generated both by the further increase in the GRP per capita level in the most prosperous (capital cities and oil producing regions) and by improving living standards in poor regions. It should be noted that there was no such picture in 1994– 1997 and, accordingly, positive changes in average indices could be triggered by marginal values.

Given the above, a gradual rise in GRP per capita over several recent years has been also characterized by a broadening gap between regions, which is evidenced by both a steady excess of the average per capita GRP value over the median one and by a growing index of standard bias of GRP in 2002. One can single out a number of regions for which the GRP per capita value appears substantially greater vs. the average one. This is true for the period between 1994 through 2002 for such regions as Nenetsky AO, Khanty-Mansi AO, Yamalo-Nenetsky AO, Altay Republic, Republic of Khakassia, Republic of Sakha (Yakutia), Tyumen oblast, and the city of Moscow.

2.2.2. Testing Hypothesis of –Convergence The concept of –convergence is accurate in the event there exists a decrease in the dispersion of the GRP per capita index for a group of regions. In other words, if t+T

Because of this, to describe an disparity between regions, one would employ the variation coefficient that is computed in the following way:

V CV = (2.10) Yavg N where V = (Yi - Yavg )2 – dispersion of GRP per capita, Yavg – N i = average GRP per capita value for regions, N – the overall number of regions.

By its structure, the index CV does not take into account a relative number of regional population, which can be considered by means of the weighted variation coefficient – CVw, which is computed by analogue with (2.10), except for the dispersion index, which is modified for the purpose of taking into account the weight of the share of the population of a single region in the total number of population ( pi ) using the following – hereinafter referred to as «Weighted Correlation Rate (method 1)”22:

N Vw = (Yi - Yavg )2 (1- pi ). (2.11) N i = There also exists another way to compute the weighted dispersion index that takes into account an average weighted value. It can be computed as follows (hereinafter referred to as Weighted Correlation Rate (method 2)”:

N Vw = (Yi - Yavg )2 pi. (2.12) i=Yet another, rather popular, index that characterize the disparity level between countries (regions) in terms of income level is Gini coefficient, which is computed by Lorenz curve. The latter is a correlation between the accumulated share of the regions’ GRP=s per capita in the total GRP and the accumulated share of the regions’ population in the overall number of population. To build this particular curve, all the regions are ranged by the increase of the GRP per capita index.

To analyze the disparity level between countries or regions in terms of the GRP index, along with coefficient Gini, economists also employ Tale index computed according to the following formula:

N Yi Yi 1 log T (1) =. (2.13) N Y Y i = See: Castro (2004).

By analogy with Gini coefficient, should Tale index equal zero, it would mean a complete equality in terms of the index concerned, while once it equals 1, – a complete disparity.

Values of the noted characteristics of disparity level computed for the index of GRP per capita level in constant 1994 prices are given in Fig. 2.2 and 2.3. Fig. 2.2 illustrates that throughout the period in question there has been no decline in the variation coefficient, while it has begun to grow roughly since 1998. In other words, since that moment the disparity between Russia’s regions in terms of the value of their GRP=s per capita has begun growing.

Coefficient of variation Coefficient of variation (weighted) - method Coefficient of variation (weighted) - method 1.1.1.0.0.1994 1995 1996 1997 1998 1999 2000 2001 Fig 2.2. Simple and Weighted Variation Coefficients of the Index of GRP Per Capita for Russia’s Regions Gini index Teil index (right axis) 0.60 0.0.48 0.0.36 0.0.24 0.0.12 0.0.00 0.1994 1995 1996 1997 1998 1999 2000 2001 Fig. 2.3. Dynamics of Gini and Teil Indices for Russia;s Regions Computed for GRP Per Capita Index The absence of a diminishment in interregional disparity in terms of GRP per capita level follows from the dynamics of Gini and Teil indices presented in Fig. 2.3. While in the beginning of the period in question there occurred a drastic single drop of disparity between regions in terms of GRP per capita (in 1995), the disparity level has been remaining practically stable for a few years afterwards. Furthermore, since 1998 onwards one noted the predominance of a trend to growth in the inter-regional disparity in terms of income level.

Summing up the research outputs, it can be argued that the concept of –convergence has failed to find a proof in the data on Russia’s regions. Our results show that the level of inter-regional disparity by the value of per capita GRP (in constant prices) does not fall. Rather, it has been growing over past 4–5 years. The analysis of various indicators of disparity has exposed similar results. The comparison of the results with conclusions drawn on the basis of an analysis of characteristics of regional income indicators in Section 2.2.1 allows to conclude that regardless of the growth in disparity, an absolute level of per capita GRP was growing nonetheless. Such a growth was noted in respect to the sample on average and among “rich” and “poor” regions.

2.2.3. Testing the Concept of the Unconventional –convergence The existence of –convergence implies a negative statistical correlation between the growth rate of income per capita and its initial level under a cross-sectional analysis of countries or regions.

A specification of the regression model finds itself determined by the kind of –convergence to be tested. Thus, of one assesses a paired regression correlation of the growth rate of income index by constant and an initial level of this indicator, it becomes possible to test the existence of an absolute convergence. But if a given equation comprise additional exogenous parameters that characterize a difference in the level of production technologies, savings rates, population growth rates, and a number of other parameters, it is the hypothesis of conditional convergence that is tested.

It is first of all necessary to analyze results of an assessment of the paired regression correlation between the value of the GRP per capita growth rate vs. 1994 (in annual terms) and the initial level of GRP in 1994, i.e. the existence of an absolute (unconditional) –convergence. Results of the assessment of the correlation are given in Table 2.2.

The table demonstrates that the concept of –convergence is applicable to Russia’s regions over the whole period in question. This is evidenced by the negative and statistically significant coefficient under the variable of the1994 GRP logarithm. So, the regions with a smaller level of their 1994 GRPs enjoyed greater rates of growth in GRP between 1994 through 2002. The convergence rate at this junction accounts for some 0.825% annually, which is a moderate indicator for such computations (for reference: in Barrow and Salai-Martin’s papers the respective rates laid within 2–3% annually).

Table 2.Results of the Assessment of Correlation between GRP Per Capita Growth Rates and the Initial Level of GRP Logarithm of GRP per capita growth rate in constant prices Explained variable (1994–2002, annualized) Number of observations Coefficient P–value t–ñòàò.

Intercept 0.561 0.Logarithm of the –0.066 0.GRP per capita Adj. R2 0.P–value F–statistics 0.2.2.4. The Impact of Budget Policy on Convergence between Russia’s Regions: Testing the Hypothesis of Conventional –convergence So, the results of our assessments do not reject the hypothesis of an absolute –convergence for Russia’s regions over the period between 1994 through 2002. More than that, the convergence rates have proved to be extremely high. That is why we are keen to examine the role played by economic policy and, particularly, interbudgetary transfers from the federal center and capital investment of regional budgets in reaching conditions of an interregional convergence. It should also be noted that since 2002 the RF Government has implemented the federal targeted program “Reduction in differences in socio-economic development of regions of the Russian Federation (for 2002–2010 and until 2015)”, however, whereas the official information on GRP over the period 2003–04 so far has been unavailable, we have no opportunity to test the hypothesis of efficiency of this program.

Box The development of the federal targeted program “Reduction in differences in socio-economic development of regions of the Russian Federation (for 2002– 2010 and until 2015)”was launched in March 2001. The program is based upon “The main guidelines of socio-economic policy of the Government of the Russian Federation over the long-term perspective” and “The action plan of the Government of the Russian Federation in the social policy area and the economy modernization over 2000–01”. The public customer of the program is the RF Ministry of economic development and trade, while the principal designer of the program is the Council for studying into production forces under the RF Ministry of economic development and trade and the RF Academy of Sciences.

The Program objective is to diminish differences in the level of RF regions’ socio–economic development; reduction in the gap by main indicators of socio-economic development between the most developed and backward regions by 1.5 times by 2010 and 2 times by 2015.

The main goals of the Program are:

• Fostering conditions of development of the regions whose socioeconomic indicators are lower than average nationwide ones;

• Creation of a favorable environment for development of entrepreneurial operations and improvement of the investment climate;