REGIONAL ECONOMIC INEQUALITY: IMPACT ON ECONOMIC GROWTH AND ITS OPTIMAL VALUE IN RUSSIA

Regional disparity can play two roles: positive and negative. From the one hand, it may become a catalyst for economic growth when the investing capital flows from the wealthy territories to poorer regions reducing business expenses on labor and spreading the market. These processes make a contribution to optimal distribution of economic resources, economically balanced territory development, forming a common market zone in the country. Thus, we deal here with the tight bonds of regions-investors with region-recipients.

From the other hand, there may be situations when regional economic misbalance becomes an impediment for investing capital movements and favors its concentration in wealthy territories of the country. In this case the regions-investors become recipients acting as a magnet for the financial and human resources of the country and regional economic inequality soars through the time. Going further we get the economically and socially depressed lands and the poorer they are the lower chances for resolving the issue. So, if the free market does not succeed in balanced territory development the central government should intervene.

Worth noting that the redundant regional disparity is not the only reason prevented free cross-regional capital flows. The hampers can be in economic bubbles in wealthy regional markets, market agents' expectations, politic preferences, administrative barriers. Anyway, the level of regional economic disparity takes a position of a separate factor of country economic growth.

 

Evaluation of economic regional disparity

In economic theory there are several methods of disparity evaluation. The wide-known one is Gini coefficient. It allows identifying the inequality level of income distribution among groups of market agents. If apply Gini index to regional disparity it can be considered as a concentration rate of domestic product, investments, labor force. In this paper the regional disparity in Russia is taken as inequality in domestic product allocation by federal district.

The choice of domestic product is supported by the following arguments. Firstly, domestic product always has the region of origin being a result of economic activities of the residents. Secondly, if take investments, we face some difficulties with their springs and different rates of return by region.

Regarding the spatial unit being figured in Gini index evaluation for the Russian Federation we prefer federal districts as sets of geographically close regions with common economic features. The territory of the country is divided into 7 federal districts (figure 1). Each of them contains a number of regions.





Figure 1 - Federal districts of Russia

Why get them? When evaluate the economic disparity of regions, eventually, we come up against the impossible issues of domestic product distribution, for instance, between Moscow (European part of Russia) and Jewish Autonomous Oblast (the Far East). The other advantage of federal district implementation is a stability of Gini index values to region consolidations. For example, when two regions of a federal district are united in one region, the Gini index for the whole country will change despite the absence of real shifts in spatial product distribution. Thus, the aggregate format of spatial performance in Gini index evaluation has got more benefits for the further analysis and making recommendations.

Based on the above arguments, evaluating the federal district inequality in domestic product^[1] the formula of Gini index distribution is the following:

 

                     (1)

 

where G is spatial Gini index; n is number of federal districts in Russia; X_i is cumulative share of population dwelling in certain federal districts with index from 1 to i; Y_i is cumulative share of domestic product distributed in certain federal districts with index from 1 to i.

It is necessary to underline that the values of X_i, Y_i make certain Lorenz curves. On the horizontal axis the federal districts are placed in ascending order of their DDP per capita. Gini index can take values from 0 to 100%. When G=0, it means the absolute parity of federal districts by domestic product per capita and G=100% describes the total economic disintegration. Table 1 shows the spatial Gini index for the Russian federal districts in 1994-2006.

 

 

 

Table 1. Spatial Gini index in Russia.

Year	Gini index, %
1994	11.6
1995	12.7
1996	14.1
1997	14.4
1998	14.7
1999	15.2
2000	18.2
2001	18.1
2002	19.7
2003	19.8
2004	20.6
2005	23.3
2006	23.0

 

The data testifies 2-time growth of spatial Gini index in 1994-2006. It is important to focus on discrete character of G dynamic, particularly, +3% in 2000 and +2.7% in 2005. The year of 2006 is the first when G goes down. Surely, there is not enough reason for much-promising outcomes, but G=23% seems to be a sort of intermediate saturation point for the Russian economy.

 

Searching for optimal value of spatial Gini index: econometric models

The optimal G values would help to get answers on some important issues as negative/positive impact of soaring spatial disparity on economic growth, how far from the optimal level the actual G values are, of course, if the optimal ones do exist. Here we pay attention to two aspects of G impact on GDP: its absolute volumes and its growth rates.

1. Spatial Gini index influence on GDP volumes. According to the work hypothesis the dependent variable in econometric model is GDP. The independent variables may include a wide range of variables, but the empirical modeling discovers the statistically significant dependence of GDP upon two characters: G and number of employees (L). Table 2 keeps the initial data for modeling of G impact on GDP volumes.

 

 

Table 2. Initial data for modeling of G impact on GDP volumes.

Year	GDP in 2006 prices, bln.rub.	Number of employees, thousand people	Spatial Gini index
1994	18,031.65	64,785.0	0.1158
1995	17,292.35	64,149.0	0.1265
1996	16,704.41	62,928.0	0.1407
1997	16,854.75	60,021.0	0.1443
1998	16,028.87	58,437.0	0.1468
1999	16,894.43	62,475.0	0.1525
2000	18,583.87	64,516.6	0.1817
2001	19,531.65	64,980.1	0.1809
2002	20,449.64	65,573.6	0.1967
2003	21,942.46	65,979.2	0.1983
2004	23,522.32	66,407.2	0.2060
2005	25,027.75	66,791.6	0.2327
2006	26,879.80	67,174.0	0.2303

Sources: Federal Statistic Service / www.gks.ru.

 

Finally, through a number of empirical modeling, we get the following non-linear regression:

 

                                              (2)

 

where Y is GDP volume in 2006 prices; L is number of employees; G is spatial Gini index; k, a, b, m, n are coefficients.

The formula (2) admits that Y=0 when G=0. It corresponds with our hypothesis: absolute equivalence makes no economic progress. If G=1, Y=k∙L^(a+b).

In linear view of the equation (2) becomes as:

 

   (3)

R²=0.99; F=50,151; DW=2.5; a=2.48; b=-1.37; m=-9.8∙10^-5; n=7.54∙10^-10; ln(k)=0.

 

In the parenthesis under the figures the t-statistic values are indicated, R² is determination coefficient, F is F-statistic, DW is Durbin-Watson test value. The level of significance is 99% for the coefficient a, m, n; for b coefficient it is 90%.

The analytical form of equation (2) gives an opportunity to express G and L points of local extremum. Find partial derivatives:  and ; solve the equations:  and . In results we get the points of local extremum G* and L*, where , . Below the certain formulas for G* and L* are placed:

 

                                    (4)

where ; ; .

                                       (5)

where ; ; .

 

The additional verification of critic points will find out whether they are of local maximum or local minimum. Especially it is actual for G*. Now go to the modeling results (table 3), their analysis and interpretation.

 

Table 3. Local extremum values G* and L* for regression Y=f(G,L).

Year	G(min)	G(max)	Gfact	L(min)	L(max)	Lfact
1994	0.1356	0.7659	0.1158	-1,251.1	66,021.0	64,785.0
1995	0.1357	0.7657	0.1265	-1,413.2	66,183.1	64,149.0
1996	0.1359	0.7656	0.1407	-1,637.3	66,407.3	62,928.0
1997	0.1359	0.7656	0.1443	-1,695.3	66,465.3	60,021.0
1998	0.1356	0.7659	0.1468	-1,736.5	66,506.5	58,437.0
1999	0.1359	0.7655	0.1525	-1,831.4	66,601.4	62,475.0
2000	0.1357	0.7658	0.1817	-2,346.8	67,116.8	64,516.6
2001	0.1356	0.7659	0.1809	-2,330.9	67,100.9	64,980.1
2002	0.1354	0.7661	0.1967	-2,626.9	67,396.9	65,573.6
2003	0.1353	0.7662	0.1983	-2,657.7	67,427.7	65,979.2
2004	0.1351	0.7663	0.2060	-2,809.5	67,579.5	66,407.2
2005	0.1350	0.7665	0.2327	-3,352.9	68,122.9	66,791.6
2006	0.1348	0.7666	0.2303	-3,302.4	68,072.3	67,174.0

 

The analysis of data in table 3 makes the following findings.

First, regarding GDP volume maximization the certain strip of effective G values is identified: (G_min; G_max). For the period of 1994-2006 there is the permanent lower line for G values: G_min=13.5%. It reflects that the actual federal district disparity less than G_min will bring about GDP decreasing. As well as the maximum level of spatial inequality is changeless: G_max=76.6%. If admitted G_fact>G_max , GDP volume would be forced to go down due to redundant G_fact value.

Second, the wideness of the strip (G_min; G_max) is about 63 percentage points that testifies about low sensitivity of the Russian economy to the spatial disparity: GDP will grow until G values are less than 76%.

If see the classic Gini index (income sharing among households), its values near 70% are typical for poorest countries. We understand that methodologically it is not fully correct to compare the spatial Gini index with the classic one. Anyway, G_max=76% is too high.

Third, the correlation of G_fact curve with G_min , G_max ones shows that in 1996 G_fact exceeded the minimum level and favored economic growth further (figure 2).

 



Figure 2. Curves of G_fact, G_min and G_max

 

Forth, the parameter L has got the strip of its effective values too. Economically, the L_min values have no sense, but from analytical point of view the maximum level of employment (L_max) is much more interesting. Worth noting that the actual number of employees has never outrun L_max.

Thus, with econometric models we set minimum and maximum levels of spatial disparity and their influence on GDP volumes. Practically, L_min is more important than L_max. The other part of our analysis is researching a spatial disparity impact on GDP growth rate (dynamic aspect). This field lets us discover new effective rates of federal district disparity.

2. Spatial Gini index influence on GDP growth rates. The econometric modeling of G impact on GDP fluctuation finds out the same regression model as (2). But the role of dependent variable is represented by annual growth rates of GDP (V), and we deal with the function V=f(G,L). Table 4 contains V values.

 

Table 4 - Real GDP growth rates in 1994-2006, percent change from previous year (chained index)

Year	GDP chain index, %
1994	87.3
1995	95.9
1996	96.6
1997	100.9
1998	95.1
1999	105.4
2000	110.0
2001	105.1
2002	104.7
2003	107.3
2004	107.2
2005	106.4
2006	107.4

Sources: Federal Statistic Service / www.gks.ru.

 

The linear regression model is the following:

 

          (6)

R²=0.868; F=14.83; DW=2.5; a=0.62; b=-1.87; m=1.03∙10^-5; n=-8.9∙10^-11; ln(k)=0.

 

The evaluated critical points G* and L* are in the table 5. Their analysis allows making some comments.

 

Table 5 - Local extremum values G* and L* for regression V=f(G,L).

Year	G(min)	G(max)	Gfact	L(min)	L(max)	Lfact
1994	-0.0350	0.2005	0.1158	55,357.2	2,187.7	64,785.0
1995	-0.0351	0.2006	0.1265	55,168.1	2,376.8	64,149.0
1996	-0.0353	0.2008	0.1407	54,940.8	2,604.0	62,928.0
1997	-0.0356	0.2011	0.1443	54,888.5	2,656.3	60,021.0
1998	-0.0358	0.2013	0.1468	54,853.0	2,691.8	58,437.0
1999	-0.0354	0.2009	0.1525	54,776.4	2,768.5	62,475.0
2000	-0.0351	0.2005	0.1817	54,485.7	3,059.1	64,516.6
2001	-0.0350	0.2005	0.1809	54,491.5	3,053.3	64,980.1
2002	-0.0349	0.2004	0.1967	54,417.1	3,127.8	65,573.6
2003	-0.0348	0.2003	0.1983	54,413.4	3,131.5	65,979.2
2004	-0.0347	0.2002	0.2060	54,406.2	3,138.7	66,407.2
2005	-0.0346	0.2001	0.2327	54,530.4	3,014.5	66,791.6
2006	-0.0345	0.2000	0.2303	54,509.0	3,035.9	67,174.0

 

First, as in the previous case we determinate the zone of optimal G values for GDP growth rates. When G value surpasses G_max=20%, it starts to hamper GDP growth. Nevertheless, it does not mean that V<1 when G=25%. The GDP growth rate will tend to be at maximum if actual G values are closest to G_max from the right or from the left side.

Second, correlating the intervals of effective G values for the functions Y=f(G,L) and V=f(G,L), we get the zone of G values that are optimal for GDP volume and its growth rates (figure 3).



Figure 3. Zone of optimal G values for GDP and its growth rate

 

Finally, the zone of optimal G values have lower and upper boundaries: 13.5% and 20% correspondently. The lower boundary is the G_min curve for GDP volumes and the upper one is the G_max curve for GDP growth rates.

Based on the above, it is possible to draw a conclusion that in 1996-2003 the spatial disparity influenced positively on production volume. Since 2004 (the year of oil prices sharply increasing) the economic inequality of the federal districts became high enough to reduce economic growth rates.

 

Estimation of economic looses caused by redundant economic disparity of federal districts

We have defined the optimal values of regional economic inequality. We see that sometimes the actual G values are not placed in the optimal zone so it is important to evaluate GDP losses due to that.

To excel it we use model (6). The equation (3) is not suitable here because in the considered interval of G values the absolute volumes of GDP have no tends to decrease. The potential GPD growth rate is evaluated on the basis of equation (6) with G=G_max (column 4, table 6). Then, we calculate cumulative growth rates to the year 1994 on the data for the actual and potential ones (columns 3 and 5, table 6).

Table 6 - GDP losses due to redundant special disparity

Year	Actual GDP growth		Maximum possible GDP growth (G=Gmax in model (7))		Economic losses due to the redundant spatial Gini index
	Annual, %	Cumulative, %	Annual, %	Cumulative, %	% GDP ([6]=[4]-[2])	Bln.rub. in 2006 prices
1	2	3	4	5	6	7
1994	87.3	87.3	107.8	107.8	-	-
1995	95.9	83.7	107.6	115.9	-	-
1996	96.6	80.9	107.3	124.3	-	-
1997	100.9	81.6	106.7	132.6	-	-
1998	95.1	77.6	106.4	141.1	-	-
1999	105.4	81.8	107.1	151.2	-	-
2000	110.0	90.0	107.7	162.8	-	-
2001	105.1	94.6	107.8	175.6	2.7	506.4
2002	104.7	99.0	108.0	189.6	3.3	647.5
2003	107.3	106.2	108.2	205.1	0.9	174.2
2004	107.2	113.9	108.3	222.1	1.1	241.6
2005	106.4	121.2	108.4	240.9	2.0	479.9
2006	107.4	130.1	108.6	261.6	1.2	296.1

 

The data in table 6 allows making the following comments.

First, the cumulative GDP growth rates (columns 3 and 5, 2006) shows that if in 1994-2006 the G values were around the optimal level of 20%, the potential GDP would surpass the actual one in 2 times.

Second, the difference between maximum possible and actual GDP growth rates (column 6 table 17) reflects missing GDP due to the different reasons. So, in 2001-2003 G_fact was not enough high (G_fact<G_min) but it was too much high in 2004-2006 (G_fact>G_max). For instance, in 2001-2006 the total losses of GDP are estimated as 296.1 bln.rub. (approximately $10 bln.).

Thus, it is found that the regional structure of the Russian economy generates extra disparity in economic development of the federal regions that put negative pressure on GDP. Considering its immobility to the changes what is the long-term vector of spatial development? The answer will follow, if analyze the certain forecasted characters in the Concept of long-term social and economic development of the Russian Federation until 2020 (Concept 2020), worked out by the Federal Ministry for Economic Development.

 

Regional disparity in long-term aspect and its impact on economic growth

The economically balanced spatial development of the Russian Federation is one the main goals in Concept 2020. The paper confirms that at present time the redundant regional disparity becomes a serious obstacle for economic growth.

Table 7 keeps the data regarding domestic product distribution by federal regions in 2006 (actual values) and forecasted one for 2020.

 

Table 7 - District domestic product in 2006 and 2020.

Federal district	DDP per capita, thousand rub. (2006 prices)		DDP per capita to the maximum level in the sampling, %
	2006	2020	2006	2020
Central	210.9	509.3	68.4	79.1
North-West	160.0	391.6	51.9	60.8
South	70.7	180.0	22.9	27.9
Volga	116.0	276.4	37.6	42.9
Ural	308.5	644.1	100.0	100.0
Siberian	122.0	302.5	39.6	47.0
Far East	150.7	383.6	48.9	59.5
Spatial Gini Index, %	23.0	21.1	-	-

Source: evaluated on data of Concept 2020 / www.economy.gov.ru.

 

It is possible to comment that the gap of six federal districts from the Ural one will be reduced by 2020. However, the development speeds by federal district are different: Central and Far East - plus 10.7% and 10.6% respectively; North-West and Siberian - plus 8.9% and 7.4% correspondently; Volga and South - plus 5.3% and 5.0% correspondently.

Of course, if realize the strategic goal of avoiding extra disparity in regional economic development, first of all, the maximum speed of overtaking should be switched on in South and Volga federal districts. Anyway, the spatial Gini index evaluated for 2020 is 21.1% and it is almost optimal value towards the zone in the figure 5. Thus, we draw a conclusion about optimal long-term government policy in spatial economic development of the Russian Federation.


^[1] Domestic product of federal district (DDP) is a sum of certain regional domestic products.

Gusev Alexander ,
28.09.2009. Views: 3082

COMMENTS