CC BY
6
UDC 330.43
Aleshin V.B. Master's student 1st course of «International Finance» faculty Financial University under the Government of the Russian Federation
Russia, Moscow Scientific adviser: Tregub I. V.
Doctor of Economic Sciences,. Candidate of Technical Sciences IMPACT OF ECONOMIC INEQUALITY ON GDP
Abstract:
The relationship between inequality and growth is an important and controversial relationship in economic theory. In this paper, besides a short overview of the economic literature, we present selected models used to verify the relationship between growth and inequality. Our findings appear contrary to our expectations and literature, showing a significant positive relation between growth and inequality. Our findings suggest the need for further investigation into this relationship, before policy decisions can be made.
Keywords: economic growth, economic inequality, Gini coefficient, income distribution, welfare economics.
1. Introduction
Increases in income inequality have both growth-promoting effects (stronger performance incentives, as well as incentives to invest in one's own human capital, to take risks, and to make investments) and growth-dampening effects (demotivating incentives, social tensions and political unrest, declines in demand as a brake on growth). Income inequality and economic growth are perhaps two of the most discussed facets of economies discussed by common people.. Growth is necessary. It allows for people to have more of the things they want, and for more people to exist comfortably; new jobs, homes, and things which people desire. Without growth, life would never truly be able to improve, anything you had extra would have necessarily come from someone else .Growth allows for the prospects of life to tend to be good, and as such it's something people, governments, and economies all seek to strive for.
On the other hand, inequality is very depressing, or at the very least an upsetting notion. The benefits one receives from the work they do, it becomes less enjoyable when one is aware that someone else would have less than them, or more. The notion is contentious, some might speak that the concept of being able to make more motivates them, while others speak of systemic inability to better one's situation as a result of where they stand now. Anger stems from aparent unfairnesses in the system, and the notion that in sufficiently unequal societies some might find it difficult to even survive. It's a discussion which cuts to the very core of the
economic philosophies which have motivated conflict throughout history..
If growth and inequality are related, determining the exact nature of that relationship would be of paramount importance. If they trend in contrast, the obvious solution would be to increase growth and decrease inequality, but if they trend together controversy is born. How much inequality would be acceptable if it could mean having much more growth? How much growth could one willingly sacrifice to become more equal? If such a relationship were to exist, there would exist a fragile balancing act that all participants would stake in.
2. Literature Review
The study of the relationship between economic growth and the inequality in the distribution of income was first considered in the later half of the twentiefth century (Kuznets, 1955.) Kuznets sought to not only look at the relation between inequality and growth, but also to define ways to measure inequailty when there existed none before. Kuznets discussed nuances of the question with regards to concepts such as aging, familial grouping, immigration, changes of industry, and urbanization; all of which he posited would complicate and define the notion of how an economy could be inequal in a time when data was more scarce and unrefined. The relationship Kuznets focused on was the ways in which growth affected inequality, and from his findings concluded that as economies begin to grow, there develops within them inequality, but as time progresses the inequality decreases; the overall trend following some inverted U-shape. This work would later motivate other studies looking at the reverse of the proposed relationship, how inequality might affect growth.
One such further study (Barro, 2000) investigated more comprehensive data in a panel format. Observing countries in this particular scale found little signifigance for a trend between inequality and growth overall, but upon categorizing countries as either high-income or low-income showed that the
trend initially starts as being negative for developing countries, but reverses to a positive relation for developed countries. This notion that th relationship might be modifiied by other factors opened up new territory with regards to the body of knowledge.
Amongst the literature of this topic there exists some degree of controversy as to whether there is a degree of certainty that the relationship between growth and inequality is negative. In one such study (Forbes, 2000,) utilizing more recent, more refined data a study approach the topic using panel estimation was conducted which aimed to contest the certainty with which the field viewed the relationship's direction. Citing ommited variable bias; such as corruption, public expenditures on health and education, or the quality of education systems as sources for systemic errors in determing coefficients, Forbes argued that country specific circumstances that were potentially too subtle to measure could be a source for the alleged
incorrect conclusions of prior studies. In considering country-specific, constant factors standard growth/inequality models were augmented with country-specific dummy variables. In order to account for economic shocks, time-specific dummy variables were also included into the model. Contrary to the previously existing literature, Forbes found that coefficients predicting the effect of inequality on growth were positive and signifigant. Discussion was made as to how much of this drastic change in results could be attributed to using better datasets, but further discussion was made with regards to what other factors could have been responsible. Forbes commented on how the data used in the study existed within a panel covering five year increments, and how in comparison most other literature treated the subject with an orientation more targeted at the long run effects of initial inequality, and how with the future prospect of more comprehensive data which more completely spans longer periods of time the exact of changing the scope of time might have on the results of such studies. The conclusions of this study stood firm against scrutiny, and reopened the debate as to whether or not inequality is good or bad with respect to associated growth.
As the field has advanced with respect to knowledge about this topic, the data required to most comprehensively understand and observe the theoretical and empirical relationships amongst these variables has continued to increase in scope and accuracy. Even amongst the most advanced models there still exists some degree of controversy and uncertainty with regards to the exact nature of the relationship. As such, nearly all progress to be made in this topic is beyond the scope of this paper; the methods and data therein, but even if true progress is beyond the reach of these tools, taking a moment to review what has been gained and applying it to more simple settings might serve to demonstrate the success this field has made since its inception. By applying new theoretical frameworks to simple panel data, the effectiveness of new frameworks can be compared to the unsophisticated frameworks used when panel data was the only available.
Country Gini Country Gini
Name Caeffics® GDP Growth Name Coçffiçent GDP Growth
(In 2005 dollars) (In 2005 dollars)
Arqentina 44.50 9.14 Lesotho 54 17 7.09
Armenia 31 07 2.20 Lithuania 33 76 1.33
Bangladesh 32 12 5.57 Malawi 46 18 6.53
Belarus 27 72 7.74 Mali 33 02 5.82
Bulgaria 35.78 0.66 Moldova 3205 7.10
Cambodia 33 55 5.96 Montenegro 2860 2.46
China 4206 10.45 Namibia 61.32 6.04
Colombia 55.51 3.97 Nepal 32.82 4.82
Costa Rica 48 10 4.95 Netherlands 2887 1.07
Czech Republic 2663 2.30 Nigeria 42 95 784
Denmark 26.88 1.63 Nonvav 26.83 0.48
Dominican Republic 47 20 8.30 Panama 51.91 5.85
Ecuador 49.26 3.53 Paraguay 51.83 13.09
El Salvador 44 53 1.36 Peru 44.92 8.45
Estonia 32 16 2.48 Poland 33 22 3.70
Finland 27 79 299 Romania 28 16 -0 94
Georgia 42 13 6.25 Sao Tome and Principe 33.87 4.51
Germany 3063 4.09 Serbia 29 65 0.58
Greece 34 74 -5.45 Slovak Republic 27.31 4.83
Honduras 53 39 3.73 Slovenia 24.94 1.22
Hungary 29.37 079 Spain 35.75 0.01
Iceland 26.30 -2.93 Sri Lanka 36.40 8.02
India 33.90 10.26 Swaziland 51.49 1.87
Indonesia 35 57 6.22 Thailand 39.37 7.81
Ireland 3206 -0 28 Tunisia 35.79 3.25
Israel 42.78 5.75 Turitev 38 79 9.16
Italy 35.52 1.71 Ukraine 24 82 4.20
Jordan 3369 2.34 United Kingdom 38 04 1.91
Kazakhstan 28 56 7.30 Uruguay 45 32 8.40
Kvravz Republic 35.41 -0.47 Vietnam 39.25 6.42
Latvia 35 27 -0.34 Zambia 57 49 10.30
For the simple regression, the dependent variable was the GDP growth rate in terms of U.S. 2005 dollars. The dependent variable was the GINI coefficient, which is a measure of income inequality within a nation. The GINI coefficient ranges from 0 (perfect equality) to 100 (maximal inequality).
This data covers 62 countries and is obtained from the World Bank. The year 2010 is chosen as the year of observation due to it having the greatest amount of observed GINI coefficients for the past 10 years. These values are found in Table 1 and a simple trend is show in Graph 1.
"Экономика и социум" №2(33) 2017
www.iupr.ru
10
For the multiple regression the following variables were selected: savings rate, labor force, GDP per capita, gross capital accumulation, human development index, mean years of education, and the amount of scientific articles written in 2010.
The savings rate as a percentage of GDP was acquired from the World Bank for the year 2010. The total size of the labor force, GDP per capita (chained in 2005 dollars), gross capital accumulation (% GDP), and the amount of scientific journals and articles were also obtained from the World Bank.
GDP Growth vs GINI Coefficent Observations - 62
The Human Development Index value for each country and their mean years of education were obtained from the 2010 United Nation's Human Development report. There descriptive statistic can be found in Table 2. The reasoning behind the inclusion of the savings rate, labor force gross capital accumulation rate, and technology level stems from Solow's model of economic growth. According to this model, these variables attempt to explain long-run economic growth and should be controlled for when acquiring the relationship between inequality and economic growth. The inclusion of mean years of education is to accommodate The Human Capital Accumulation theory which describes the theoretical mechanism through which unequal societies, certain individuals are unable to undertake investments in their education (human capital). As such, the growth benefits of said capital are lost. HDI is included due to the Kuznetsova's theory that inequality would decrease as countries become more developed. Income has been included in prior literature to investigate the relationship between growth and inequality. This collinearity is mostly likely the result of diminishing returns of GDP per Capita on GDP growth and Kuznetsova's theorized relationship between inequality and income.
Table 2: Descriptive Statistics
"3K Variable Observations Mean Standard Deviation Minimum Maximum
GDPGrow2010 62 4.28 3.60 -5.45 13.09
er> 17 At o n ~>A QO «1 n
Figure 1: Collinearity Test
GDPGrow2010 Gini2010 PerSave2010 LabForce2010 GDPperCap2010 GrMsCaßPerc HOI MeanEökicatjon TechLevel
GDPGrow2010 1.00
Gini2010 0.45 1.00
PerSave2010 0.31 •0.03 1.00
LabForce2010 0.33 0.02 0.54 1.00
GDPperCap2010 -0.46 -0.40 0.06 -0.12 1.00
GrossCapPerc 0.35 -0.02 0.49 0.49 -0.34 1.00
HOI -0.50 -0.42 -0.05 -0.13 0.72 -0.36 1.00
McanEducation -0.47 -0.39 -0.10 -0.21 0.56 -0.37 0.87 1.00
Techlevel 0.08 -0.06 0.36 0.72 0.27 0.21 0.23 0.11 1.00
For a simple regression and multiple regression to be useful, it must satisfy all 5 of the Gauss Markov assumptions4. Since the model is written in a linear parameter, the first assumption is satisfied. The assumption of random sampling isn't satisfied due to the data only being collected on countries that allowed the World Bank researchers to visit. The strictness of this assumption can be neglected in this case. The third assumption of no perfect collinearity is satisfied with Figure 1,which shows that no independent variables have a perfect linear relationship. The fourth assumption of zero conditional mean cannot be proven, while it will be assumed in order to perform these regression, the current datasets are still lacking to some degree in completeness and rigorous definition; as such we can not certainly say that this paper meets assumption 4. Finally, the assumption of homoskedasticity is satisfied because the error term is assumed to have the same variance given any values of the explanatory variables.
4. Results
4.1 Simple regression
To find the relationship between income inequality and GDP growth, first
4 I.V.Tregub. Mathematical models of economic systems dynamics: Monography. M.: Finance Academy, 2009. 120 p.
we did a simple regression, setting 2010 GDP growth as a dependant variable and 2010 GINI as an independent variable. Equation is : GDPGrow2010 = fi0+ fi1(Gini2010) + u
To test the significance and effect of GINI on growth, the t-value, p-value and R2value was also calculated and shown in table 3 below.
Table 3: OLS Simple Regression]
Dependent Variable: GDPGrow2010
Independent Variable Coefficient t-value p-value
Gini2010 .1771*** 3.89 0.00 2017
Constant -2.36 -1.34 .184
N=62. 'Significant at 10%, **5%, ***!%
Having t value of 3.89 or p value of 0.00, our result is statistically significant. If we plug our findings into our equation, result would look like: GDPGrow2010 = -2.36 + .1771 (GM2010)
From here, we can see that in contrary to our initial hypothesis, we got positive relation between GINI and GDP growth. Regression coefficient of GINI was 0.1771, which means as GINI increase by 1 point, GDP growth rate also increases by 0.1771%. Our findings suggests that if we have ultimate equality, GDP growth would be -2.36% which results decrease in GDP. Based on our results, we can aslo interpret that in order to have growth of GDP, GINI should be more than 13.33.
4.2 Multiple regression.
To make our finding more accurate by avoiding omited biases, we did a multiple regression test with more independent variables that are mentioned above in data section of this paper. Our multiple regression equation is:
GDPGrow2010 = fi0+ fi1(Gini2010) + fi2(PerSave2010) + fi 3(LabForce2010) + fi4(GDPperCap2010)+ fi5(GrossCapPerc) + fi6(HDI) + fi 7(MeanEducation) + fi8(TechLevel) + u
Like we did in single regression calculation, the t-values, p-values and R2 value was also calculated and shown in table 4 below.
Table 4: OLS Multiple Regression
Dependent Variable:GDPGro\v2010
Independent Variable Coefficient t-value p-value RA2
Gini2010 0.12** 2.45 0.02 0.4527
PerSave2010 0.09* 1.77 0.08
LabForce2010 3.26e-09 0.53 0.60
GDPperCap2010 -4.24e-05 -1.22 0.23
GrossCapPerc 0.02 0.31 0.76
HDI -3.90 -0.63 0.53
Mea^ducatjon -0.06 -0.18 0.86
TechLeyd 6.81e-06 0.14 0.89
Constant 1.12 0.27 0.79
Our main independent variable Gini2010 is statistically significant at 5% and PerSave2010 is statistically significant at 10%, whereas other independent variables are statistically insignificant. our equation would look like below, after plugging results.
GDPGrow2010 = 1.12 + .12(Gini2010) + .09(PerSave2010) + 3.26e-09(LabForce2010) + -4.24e-05(GDPperCap2010) + .02(GrossCapPerc) + -3.90(HDI) + -. 06(MeanEducation) + 6.81e-06(TechLevel) Again, oppose to our hypothesis, multiple regression model suggests positive relation between GDP growth and GINI. If country has ultimate income equality, GDP growth would be 1.12% and country sees 0.12 percentage point increase in GDP growth as a consequence of 1 point increase in GINI. GDP would increase by 0.09% when savings rate increases by 1%. Another interesting finding is mean years of education and human development index have negative relation with GDP growth. 0.01 point increase in HDI and 1 years of more education causes 0.039% and 0.06% decrease in GDP growth respectively, but these variables are statistically insignificant.
Since the independent variables beside Gini2010 are not quite significant, we did the F test hypothesizing all those 7 variables are equal to 0. Because this hypothesis leaves us with only one independent variable, our single regression model become our restricted model. If we calculate F value by using R2values from both model and restricting those 7 variiables, F value equals to 3.47 which is statistically significant. Therefore we can't accept the robustness test's null hypothesis. In other word those 7 variables have joint effect on our dependent variable GDP growth and it also means our multiple regression model is better that our single regression model int terms of accuracy.
After determining better model, we tested our initial hypotesis as a one-sided t-test. For all signifigance levels, since our observed coeffieicents were positive they failed to surpass the critical values for our hypothesis, as such we failed to reject the null hypothesis that the coefficients are not negative.
5. Conclusion
Our hypothesis of a negative relationship between economic growth and inequality was not corroborated by the simple and multiple model regression that were run. Both models showed a positive relationship to economic growth with a 1 point increase in GINI causing growth rate to increase by .12%. Counter intuitively, an increase of 1 year of mean education decreases growth rate by.06%; while this result is found to be insignifigant, it still is of interest to note the counterintuitive nature of this result. This result not only challenges our hypothesis and the literature supporting it, but it also implies troubling policy implications if our results are true.
The notion that something like growth, which is considered universally useful and beneficial to society would potentially have a positive relationship with something as troubling as inequality leads to a natural conundrum. If we want to increase growth, are results indicate that there could be an associated increase in inequality as a direct result of that increase in growth. If we wish to decrease inequality, our results indicate that there could be an associated loss of growth that results from this decrease in inequality. The implications of these findings suggest that there exists some trade-off between growth and inequality which must be made whenever one wishes to alter one or the other. If this is a case, the question of determining what the acceptable level of inequality is, and how much growth we need, and attempt to consolidate those two requirements. The policy implications of such findings would further complicate and already contentious issue.
While the results of this paper seem grim, perhaps it is premature to speak with definitiveness. Data regarding inequality is often underreported, and the question of which forms of inequality are most influential on the relationship is a question rarely addressed even now. Further questions are posed by how our findings clearly represent short-term relationships, while more comprehensive series investigate long and medium-run relationships between past inequality and current growth. Further questions are posed by the issues of collinearity in the multiple regression model, and our own uncertainty of whether omited variable bias could yet be involved. With these ascpects of uncertainty, it seems that further investigation into this relationship is required, more comprehensive and longer term data will continue to become availiable, and with it a more thorough understanding of the relationship will be possible.
References:
1. Kuznets, Simon. "Economic Growth and Income Inequality." The American Economic Review45, no. 1 (1955): 1-28.
2. Barro, Robert J. "Inequality and Growth in a Panel of Countries." Journal of Economic Growth5, no. 1 (2000): 5-32
3. Forbes, Kristin J. "A Reassessment of the Relationship between Inequality and Growth." American economic review(2000): 869-887
4. World Bank Open Data (http://data.worldbank.org/)
5. I.V.Tregub. Mathematical models of economic systems
dynamics:Monography. M.: Finance Academy, 2009. 120 p. УДК 336.71.078.3
Al Khasan I.D. master student of 1st year IFF group 1-2m Scientific advisers: Tregub I. V.
Ph.D. in Technics, Sc.D. in Economics, Professor
Fedunin A.S.
Ph.D. in Economics, docent Financial University under the Government of the RF
Moscow, Russia
Аль Хасан И.Д. студент 1 курса магистратуры МФФ группа 1-2м Трегуб И.В., к.тн., д.э.н. научные руководитель, профессор Федюнин А.С., к.э.н. научные руководитель,, доцент Финансовый университет при Правительстве РФ
Россия, г. Москва DETERMINATION OF ADDITIONAL BANKING STABILITY REGULATION FACTORS TO BASEL III STANDARD ОПРЕДЕЛЕНИЕ ДОПОЛНИТЕЛЬНЫХ ФАКТОРОВ РЕГУЛИРОВАНИЯ БАНКОВСКОЙ СТАБИЛЬНОСТИ К СТАНДАРТУ БАЗЕЛЬ III Annotation: Banking regulation is very complex process which is regulated basing on international standards such as Basel standards with the newest Basel III. However, there may be some additional elements not considered or weakly controlled by Basel III. In this article such factors are considered and explained using econometrics methods.
Key words: Basel III, Econometrics, Banking, Regulation Аннотация: Банковское регулирование - очень сложный процесс, который регулируется на базе международных стандартов, таких как стандарты Базель, где Базель III является новейшим стандартом. Однако могут существовать дополнительные элементы, которые не рассматриваются или слабо контролируются стандартом Базель III. В этой статье рассматриваются и объясняются такие факторы, используя эконометрические методы.
Ключевые слова: Базель III, Эконометрика, Банковское дело, Регулирование
