Bachelor thesis
the effect of income inequality on the gross domestic product growth
rate
28-06-2016
Frits Eringa (10545204)
Supervisor: E. Westerhout
Study programme: economics and finance
Number of credits: 12
Form for thesis
Your name: Frits Eringa
Your student number: 10545204
Specialization (within Economics and Business): Economics and
Field: macroeconomics, income inequality
Number of credits thesis: 12
Title of your research proposal: the effect of income inequality on the gross domestic
product growth rate
Assigned supervisor (to be filled in by thesis coordinator):
If a teacher has already accepted to supervise your thesis, please provide the name.
Name of supervisor: Ed Westerhout
Statement of Originality
This document is written by Student Frits Eringa who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Introduction
The last few decades have been accompanied by some astounding accomplishments surrounding poverty and inequality; the latest estimates, from the World Development Indicators 2015, show that the number ofpeople living in developing countries under the poverty line of $1.25 per day fell from 43.6% percent of the population in 1990 to 17.0% in 2015. This drop is mostly accounted to the substantial economic growth in Asian regions. Affluent societies, such as those in Europe, have maintained their reduction in extreme poverty. Which means that they started off with low poverty rates and maintained their reduction through the mid 1990’s to reach the target by 2010.
One of the main forces to drive down poverty is macroeconomic growth. Most theoretical models predict that economic growth will be accompanied with a raise in labour demand and thus raising wages, this will lead to less income inequality and so reduce poverty, or so the models predict. However, economic growth does not necessarily mean a drop in poverty. In reality great economic growth is often accompanied by an increase in poverty. One of the reasons for this is that the initial distribution of capital is not equally divided amongst a population and since the return on assets is generally relatively high an increase in economic growth will only expand the wedge between the rich and the poor.
To reduce the widening of the gap between the rich and the poor governments implement social policies which aim to redistribute the newly acquired wealth among the impoverished in a population.The dominant opinion in western civilizations is thatredistributive policies are therefore a favourable instrument to combat extreme poverty and income inequality.However the effects of redistributive policies are still being debated by scholars and therefore are often identified as
it will hamper economic growth or whether it will spur it. Since many countries aim to reduce income inequality and the effect of this reduction in GDP growth is still not clearly defined the research question of this paper and a relevant question in modern day economics is: what is the effect of income inequality on the GDP growth rate?
A large body of literature has already been written about the link between GDP growth and income inequality, and since the release of Capital in the Twenty-First Century by Thomas Piketty this debate has found renewed attention by economist around the globe. In this book Piketty argues that the return on capital in developed countries is consistently higher than the overall economic growth in these countries, thus the gap between the affluent and the poor will only increase. However other scholars argue that inequality is beneficial for economic growth since it raises incentive to increase efficiency and labour supply. Altogether it seems that there still is turbulence surrounding this subject, this paper seeks to clarify the nexus between income inequality and GDP growth, as to give policy makers a clear view of the implications of their policies. It will also briefly touch on some of the underlying factors driving this relationship and will provide some personal insight as how to handle this problem.
The first section of this paper will discuss the current literature and compare several scholarly opinions with each other, the results of their papers will be discussed in depth and by doing so it will be analysed what has already been discussed about this subject.
Then the methodology of this paper will be explained, the chosen data will be justified and their origin, efficiency and robustness will be discussed. A hypothesis will be formulated with the expected results of the data and after this is done the collected data will be presented, analysed and compared with the hypothesis.
Finally some concluding remarks will be given on the total of this paper.
Literature review
Income distribution and the inequality that often accompanies this subject have always been a principal topic of debate among researchers and economists. Simon Kuznets (1955) has initiated the first empirical research regarding this subject when he published his paper which introduced the Kuznets-Curve; this curve describes inequality in relation to income per capita. The Kuznets curve takes a hump-shaped form, which implies that for low levels of economic growth inequality rises, but as income per capita starts to rise inequality will fall. However, in his concluding remarks Kuznets
(1955) noted that the paper is about 5 percent empirical information and 95 percent speculation. He added that more empirical research is needed since income distribution is a focal point at which the functioning of an economy depends. Furthermore, the theory seems to imply that the movement of growth and inequality is initiated by a shift in the working population; people from the agricultural sector, which is assumed to have a low mean income and low inequality, move to a more
industrialized working environment, which has a higher mean income but also higher inequality. Later the inequality is reduced, by, for example, cushioning policies which benefit the poor, but the mean income stays at the higher level. It is arguable if this movement in working population is still relevant for developed economies.
Even though a lot of theoretical models can accommodate the Kuznets-Curve there seems to be little empirical data whichsupport the theory, Deininger and Squire (1996) sought out to test the theory against high-quality empirical data. they conducted study in which they analysed over a 100 countries in developed as well as developing economies. The aim of the study was to provide the theoretical framework, which implied a systematic link between inequality and growth rates, with empirical data to back it up. They took data from micro-level household surveys and supplemented this with country-level data over a time period of fourteen years. Their research yielded results that contradict the Kuznets-Curve; they found that whether average income was increasing or declining, the effect of the Gini Coefficient, one of the widely used instruments to measure income inequality, was not significant. To illustrate this look at the change in real income of the US, from 1950 to 1991 it rose from $8,772 to $17,594 while the Gini Coefficient moved from 36.0 to 37.9. Even in the cases where the Gini Coefficient did move substantially, as in Thailand where it rose from 41.3 to 51.5, the change seems small when compared to the rise in real income, which fourfold in the same period of time (Deininger and Squire 1996).
Interestingly enough Lin, Huang & Yen sought out to find the threshold of income for which inequality will have a positive effect on growth. They argue that inequality does not need to have either a strictly positive effect or a strictly negative effect, but that this relationship might reverse when exceeding a certain threshold. This study focusses on state-level differences, namely in the United States, instead of a cross country analysis. This has several advantages, firstly all the data is highly comparable since most information is acquired for nationwide surveys taking the same form and time period. Secondly there will be no country specific deviations among the outcomes since states are generally more homogenous than countries are. The sample taken are 48 states over a time period of 59 years. two regression are performed, one linear and one nonlinear, the latter being used to locate that threshold for which the marginal effect of inequality on growth changes
lower growth, this seems to correspond with the results of Murat, Iyigun & Owen. This effect recedes when income rises and eventually even reverses itself. So when an economy is highly developed the nexus between inequality and GDP growth is significantly positive according to this data. (Lin, Huang & yen, 2012)
Later Deininger and Squire (1998) researched a comparable phenomenon, only this time taking economic growth as the dependentvariable, and the initial income inequality of a country as the independent variable. They do not necessarily imply a humped shaped curve but they do analyse the effect of income inequality on GDP growth. This is more interesting for this research paper since the same relation will be studied. The complete regression formula looked as follows:
𝐺𝑟𝑜𝑤𝑡ℎ𝑖𝑡 = 𝑎 + 𝑏 𝐼𝐺𝐷𝑃𝑖𝑡+ 𝑐 𝐼𝐺𝐼𝑁𝐼𝑖𝑡+ 𝑑 𝐼𝑁𝑉𝑖𝑡+ 𝑒 𝐵𝑀𝑃𝑖𝑡+ 𝑓 𝐸𝐷𝑈𝑖𝑡+ 𝑒𝑟𝑟𝑜𝑟
where i denotes countries, t denotes time, IGDP denotes initial GDP, IGINI is a measure of initial income inequality, INV indicates investment, BMP represents the black market premium, and EDU is education as measured by either average attainment in the population or enrolment rates (Deininger and Squire 1998). First a regression was run by using the average of high-quality observations on income inequality, the finding was that initial income inequality indeed negatively affects GDP growth. However when regional dummies were added to the regression the results of the initial income inequality ceased to be significant. Which leads us to question the robustness and validity of the negative relationship between income inequality and GDP growth. Therefore a second regression was run where initial distribution of land was used as an independent variable, the findings were again highly significant. Even when the regional dummy variables were added there was still a significant negative link between initial distribution of land and the GDP growth. This might suggest that
collateral related constraints limit the poor of a society to credit markets and thus drive down the GDP growth. Likewise, the possession of land may prove to be a factor which greatly enhances productivity, especially in developing economies where agriculture still dominates the market. However, data on land inequality was very limited and it could not be used in the panel data model to check if cross sectional results hold after controlling for omitted variable bias. (Tabassu & Majeed, 2009)
Cornia & Court (2004) have published an extensive report on inequality growth and poverty, which also discusses the effects of inequality on the GDP of an economy. They suggest that there might be an optimal level for inequality, which means that there is a level of inequality which where economic growth reaches its peak, which varies for different economies as they have different structures. Too low levels of inequality might be detrimental for growth since it is often accompanied by a very compressed wage distribution. This leads to wages which are too low in many cases and therefore do not accurately reward different capabilities and thus reduce a great incentive for
workers, since they are not accurately rewarded for their capabilities. They also increase labour-shirking and free-riding. This means that having too low income inequality hampers the efficiency of labour markets. All these effects reduce the GDP of an economy. This also occurs when the
redistributive polices that are in place are too high, such as a too high marginal tax rate. On the other hand there are economies where the gap between the rich and the poor is excessively wide. This is also harmful for an economy since it encourages rent-seeking and predatory behaviour for the asset-holders and it decreases the work incentive of the asset-less. Especially unequal rural economies tend to be less efficient when compared to more equal rural economies. Furthermore high inequality in assets can hinder progress of education and human capital accumulation. When the distribution of asset holders is unequal in a country it can also have an effect on policies which are suboptimal, Cornia & court (2004) postulate, for example, a theory that inequality in democratic societies can lead to populistic policies which have a negative effect on efficiency, growth and stability. Finally inequality can lead to social tension which also hampers the GDP growth, it can erode the security of property rights, augment the threat of expropriation and drive away domestic and foreign investments. Inequality is also closely correlated with high crime rates. To mitigate to effect that harm growth and promote the effects that are pro-growth policymakers should target the ‘efficient inequality range’, this range should be between the values 25 and 40 of the pre-discussed Gini coefficient. It should be noted that this range might vary for different economies with vastly different resources.
Tabassum & Majeed (2009) also have studied the effect of income inequality on GDP and also included the role of credit market imperfections in their regression. In this research the Gini
Coefficient is again used to measure income inequality, however since missing data was quite frequent regarding this coefficient they decided to supplement it with household surveys on poverty and inequality from several organisation such as the World Bank and the IMF. Nonetheless the data still was not entirely accurate since these surveys differed across countries and were not perfectly comparable with each other. This has to be considered when analysing the results. To measure credit market imperfection a dummy variable was added which takes the value one when a country has a level of financial intermediation which is above the sample median. The level of financial
intermediation is measured by the share of broad money (M2) as a percentage of the GDP and the share of credit to the economy in GDP (Tabassum & Majeed, 2009). the results of the regression showed that there is a highly significant negative link between income inequality and economic growth, especially in the long run. This effect is intensified by credit market imperfections which further declines the growth rates, it is argued that one of the main reasons is the inability for a major part of a population to invest in human or physical capital due to these imperfections on the credit market. Some of the empirical data suggest that in the short run the effect of income inequality might
be positive, but in the long run, due to credit market imperfection the effect is strongly negative. A reduction in inequality will therefore raise the quantity and average productivity of investment and is therefore beneficial for macroeconomic growth. The results that this study yielded where highly significant, except for the short time period impact of investment in physical and human capital on growth rate, on the long run these results were significant.
Murat, Iyigun & Owen (2004) have done a similar study, the focal point of this research, however, was focussed on the initial level of income per capita. That is whether an economy is developed or not and link this to the variability in aggregate consumption and aggregate GDP growth rates. It must be noted that variability in GDP growth is not entirely the same as GDP growth, it is however interesting to study the effect to which extend the GDP might differ for certain levels of income inequality. They have used four different instruments to measure income inequality; the Gini Coefficient and three other measures which divides the population into different income classes which earn a certain percentage of the total income. The number of observations was about 40 which leads a few of the results to be insignificant, some conclusions can still be drawn from this however. When analysing the regression results it can be concluded that in developing economies, id est economies with a low income per capita, greater inequality results in less variability in GDP growth. Effect is reversed when looking at economies with a high income per capita, here greater inequality leads to a higher variability in GDP growth. Still, as aforementioned, these results were not always as significant as should be, especially the robustness was inadequate when using alternative sample selection strategies and estimation techniques ( Murat, Iyigun & Owen, 2004).
This leads us to discuss and analyse the different channels through which income inequality affects the GDP growth rate. As argued above the dominant effect on which the relationship between growth and inequality depends may differ for different states of the economy. For developing
economies, which have a low real initial income per capita, the nexus between GDP and inequality is negative. This effect is attributed by Galor and Zeira (1993) to imperfect capital markets; this limits the poorer individuals of a society from fully utilising their human or physical capital due to credit
restraints. This effect not only limits itself to the short run, but it seems to influence the long run as well. Thus growth is affected by the initial distribution of wealth among a society, to be more precise, it is determined by the percentage of people who have inherited sufficient wealth to invest in human capital. The researchers end there remarks by concluding that it is beneficial for the growth in the long and the short run to have a large middle class.
Then we have developed economies, here the theory of innovation and efficient market allocation seems to prevail. Bhattacharya (1998) argues that higher inequality will lead to more physical capital accumulation with high rates of return and so a higher GDP. The study identifies a
portion of an economy as capitalists, who are assumed to have access to risky assets with high rates of returns. These capitalists can finance their projects internally and thus minimise the standard costly state verification (CSV) problem. Thus an reduction of income inequality will only serve to lower the steady state value of capital stock and so lower the overall growth of an economy. This also means that measures taken by policymakers to redistribute wealth might have a maleficent effect for all the agents functioning in the economy. It must be noted however that this research has been restricted to steady state values, so for an economy with dynamic equilibria, which is often seen as a more realistic option, these results might not always hold.
Furthermore Bell and Freeman (2001) point out that high inequality in pay gives workers an incentive to work longer and more efficient, also it will allocate resources to relatively unequal areas. The study focussed on the United States and Germany, where the workers of the former country see a greater wedge in pay, but also work more hours. The researchers set out to find a relation between inequality and hours worked and they conclude that the data is consistent with the theory that greater inequality generates additional hours worked. They have also shown that greater hours worked improves future earnings and the chance of being promoted, which in turn leads to more hours being worked; since it is then more costly for workers to not work. Logically, when a population of an economy starts to work more hours, this will raise the level of GDP. Hence the conclusion that higher inequality yields a higher GDP growth rate.
Lastly a recent OECD paper has been published by Cingano (2014), he argues that the gap in the OECD countries between the rich and the poor has been continuously growing since the 1980s . this effect cannot be solely attributed to a rise in the top incomes but also due to the slow growth of the middleclass, which is troublesome according to him. The results of this paper find that a growing gap between the low and middle income household and the rest of the population is most harmful for GDP growth and statistically significant. However, there was no evidence found that the further enrichment of those with high incomes is detrimental for an economy. The harmonised data is drawn from solely OECD countries over the past thirty years and it also analyses the effect of redistributive policies on long term growth; Cingano (2004) found that these sort of policies, which succeed in achieving greater equality in disposable income amongst a population, are definitely not adverse to growth. He found that one of the main reasons why great inequality reduces growth is because it prevents a large segment of the population from investing in human capital, more specifically adequate education.
To obtain the data required for this paper the gross domestic product growth rates from OECD countries will be used. This is used as the dependent variable. Furthermore this data is utilized since the GDP of the previous time period, defined as t-1, is likely to have an influence on this current growth rate. The data will be obtained from The World Bank and will be the annual percentage growth rate of the gross domestic product at market prices based on the local currencies. The gross domestic product is the sum of all value added to an economy of a country plus any taxes and minus subsidies, it is calculated without accounting for depreciation of assets or the depletion and degradation of natural resources.The time period chosen will be from 2004-2013, this relatively small time period is chosen because data concerning the Gini coefficient is not widely available for dates before 2004. This paper focusses on high quality primary data, all acquired from the same source, therefore the time period and number of countries is limited. the analysis that will be performed will take the form of a cross-country analysis including the following fourteen OECD countries: Belgium, Canada, Czech Republic, Estonia, Finland, Greece, Iceland, Ireland, Italy, Luxembourg, Poland, Portugal, Slovak Republic and Slovenia.
To measure income inequality the Gini coefficient will be utilized. This coefficient calculates the degree to which a country deviates from perfect equality. This is done by first drawing a Lorenz curve, which plots the cumulative percentages of total income received against the cumulative total number of recipients, beginning from lowest to highest. Then a perfect equality line is drawn, this is a line of 45 degrees which starts in the origin. The Gini coefficient is the result of the area between the perfect equality line and the Lorenz curve divided by the total area under the perfect equality line. This always yields an Gini coefficient between 0 and 1, or 0 and 100 when expressed as percentages; 0 means perfect equality while 1 or 100 implies perfect inequality. This coefficient is chosen since it is one of the most popular estimates to calculate income inequality. Furthermore the Gini coefficient will be squared since I hypothesize that the relation between GDP and income inequality will be hump-shaped. A problem however is missing data, for most of the countries the data is incomplete. Thus a small dataset will be used in this paper. The information is collected from The World Bank.
After this simple regression is run, some variables will be added to see whether this changes the coefficients or the significance. Firstly the GDP from a year before is added, because I suspect that this influences that current GDP greatly. The regression so far will have to following form:
𝐺𝐷𝑃𝑐𝑜𝑢𝑛𝑡𝑟𝑦= 𝛽0+ 𝛽1 𝐺𝐼𝑁𝐼𝑐𝑜𝑢𝑛𝑡𝑟𝑦+ 𝛽2 𝐺𝐼𝑁𝐼𝑐𝑜𝑢𝑛𝑡𝑟𝑦2 + 𝛽3 𝐺𝐷𝑃𝑐𝑜𝑢𝑛𝑡𝑟𝑦, 𝑡−1+ 𝜀𝑐𝑜𝑢𝑛𝑡𝑟𝑦
furthermore a country’s social policies will be analyzed, to be precise: a country’s subsidies and other transfers as a percentage of total expense, this includes grants, other social benefits to private and public enterprises and social security, social assistance benefits and employer social
benefits. The data required for this analyses is also retrieved from The World Bank. This will yield useful information on whether social policies that promote inequality hamper or benefit growth. Also, the assumption is drawn that a country with more social policies can be seen as an more developed economy than an country with less social policies.
The GDP is also affected by the trend of the business cycles, when the economy is in an recession the GDP will drop regardless of the level of income inequality in a country. Therefore a variable has been added which will compensate the GDP when the economy is in a slump at a certain time and decrease the GDP when the economy was booming. Whether the economy is expanding or in a recession will be measured by using the Composite Leading Indicators. These are indicators that show fluctuations of economic activity, this data is acquired from The Organisation for Economic Co-operation and Development. Lastly and error term is added to account for any measurement errors. When all this data in combined it is possible to perform a final regression analysis which has the following form:
𝐺𝐷𝑃𝑐𝑜𝑢𝑛𝑡𝑟𝑦= 𝛽0+ 𝛽1 𝐺𝐼𝑁𝐼𝑐𝑜𝑢𝑛𝑡𝑟𝑦+ 𝛽2 𝐺𝐼𝑁𝐼𝑐𝑜𝑢𝑛𝑡𝑟𝑦2 + 𝛽3 𝐺𝐷𝑃𝑐𝑜𝑢𝑛𝑡𝑟𝑦, 𝑡−1+ 𝛽4 𝑆𝑂𝐶𝑐𝑜𝑢𝑛𝑡𝑟𝑦
+ 𝛽5 𝐸𝐶𝑂𝑁𝑐𝑜𝑢𝑛𝑡𝑟𝑦+ 𝜀𝑐𝑜𝑢𝑛𝑡𝑟𝑦
Where GDP is the gross domestic product growth of a country, 𝛽0 is a predefined constant, GINI is the
Gini coefficient, SOC measures the number of social benefits as a percentage of total expenses, ECON is an variable that estimates that effect of the current state of the economy and 𝜀𝑐𝑜𝑢𝑛𝑡𝑟𝑦 is an error
term.
Hypothese
I expect that the effect of income inequality on GDP will be hump-shaped, with this I mean that a certain level of inequality is desirable when promoting growth in an economy, however when inequality starts to rise the growth will decelerate. I even suspect that when a certain threshold is exceeded that inequality will have a negative effect on macroeconomic growth. I expect that this relation will be hump-shaped because a certain level of inequality is useful since this will promote investments in assets with relatively high returns. However when inequality becomes excessive a considerable part of the population will not be able to invest in human capital, I suspect that this effect will eventually become so large that is dominates the initially positive effect of the higher returns on certain assets. Furthermore I suspect that the relationship for developed countries, which I define as countries that have an abundance of social policies, will be predominantly positive. By cause of an increase in efficiency and positive incentives this effect will be mostly constructive for an
economy. Thus for developing countries I suspect that higher levels of inequality will be
disadvantageous for the GDP growth, because the effect of imperfect capital markets will prevail. Data analysis
In this section two separate regressions will be ran for all the different countries, the results have been extracted and put in the table below for clarity. The first regression will have to following form:
𝐺𝐷𝑃𝑐𝑜𝑢𝑛𝑡𝑟𝑦 = 𝛽0+ 𝛽1𝐺𝐼𝑁𝐼 + 𝛽2𝐺𝐼𝑁𝐼2+ 𝛽3𝐺𝐷𝑃𝑡−1+ 𝜀𝐼
Since for most countries, due to limited data, a small sample was drawn many results are deemed insignificant. Results that are slightly significant* have been highlighted.
*note: slightly significant has been interpreted as having a p-value around 0.25
The table contains the coefficient of the different beta’s, it is in brackets if the number is negative. Behind the coefficient the corresponding p-value is in brackets. Even though most of the results are insignificant, some conclusions can still be drawn from this analysis. Firstly we can see that in every single country the relationship between inequality and GDP is a hump-shaped curve, because the sign of the GINI and GINI2 coefficient changes for every country. It is also interesting to note that for six of the fourteen countries, which have a negative GINI coefficient but a positive GINI2, for higher levels of inequality the effect on GDP will positive, for low levels of inequality the effect will be
reversed. For the other eight countries the GINI is positive and GINI2 is negative, which shows that for high levels of inequality the effect on GDP will be negative, but for low levels of inequality the effect
B_0
B_1
B_2
B_3
R2
N
Belgium
960,91 (0,58)
(7038,56) (0,57)
12903,30 (0,57)
(0,01) (0,98)
-0,15
9
Canada
12542,80 (0,40) (78409,11) (0,40) 122556,60 (0,40)
(0,27) (0,47)
-0,13 11
Czech rep.
(3755,61) (0,53) 28131,00 (0,53)
(52627,69) (0,54)
0,62 (0,14)
0,19
9
Estionia
(1903,19) (0,59) 11160,60 (0,60)
(16316,82) (0,61)
0,42 (0,32)
0,12
9
Finland
(1951,24) (0,73) 14133,61 (0,75)
(25507,80) (0,757) 0,03 (0,91)
0,27 12
Greece
(9775,94) (0,13) 57771,76 (0,13)
(85344,82) (0,129) 1,10 (0,01)
0,71
9
Iceland
(349,92) (0,54)
2547,39 (0,53)
(4598,80) (0,52)
0,48 (0,53)
-0,2
9
Ireland
34,81 (0,99)
(349,70) (0,98)
774,58 (0,97)
0,47 (0,26)
0,03
9
Italy
492,20 (0,89)
(3203,39) (0,88)
5197,69 (0,88)
0,06 (0,91)
-0,41
9
Luxembourg 139,23 (0,88)
(734,19) (0,91)
904,13 (0,94)
0,24 (0,61)
-0,12
9
Poland
(193,15) (0,23)
1146,20 (0,23)
(1643,45) (0,23)
0,01 (0,99)
-0,03
9
Portugal
(949,38) (0,01)
5183,27 (0,01)
(7053,82) (0,01)
(0,44) (0,10)
0,81
9
Slov. Rep.
950,81 (0,25)
(7026,48) (0,25)
13005,44 (0,25)
(0,00) (0,99)
-0,15
9
Slovenia
(2075,82) (0,62) 17978,26 (0,607) (38766,46) (0,59)
(0,45) (0,38)
0,25
9
on GDP will be positive. This evidence might suggests that the relationship of income inequality on GDP may vary for different countries. These results might be attributed to different levels of
development in which these economies are, however since most of the countries are a member of the European Union this is not highly likely since these countries are all regarded as developed countries. Another reason for this difference might be access to credit markets which disables some of the poorer people of a society to adequately invest in capital, most importantly human capital.
It must be noted however that some of the drawn conclusion are based on speculation, since the robustness of this data is questionable and the sample size is limited.
The second regression has the following two control variables added which gives it the following form: 𝐺𝐷𝑃𝑐𝑜𝑢𝑛𝑡𝑟𝑦= 𝛽0+ 𝛽1𝐺𝐼𝑁𝐼 + 𝛽2𝐺𝐼𝑁𝐼2+ 𝛽3𝐺𝐷𝑃𝑡−1+ +𝛽4𝑆𝑂𝐶 + 𝛽5𝐸𝐶𝑂𝑁 + 𝜀𝐼
Since for most countries, due to limited data, a small sample was drawn many results are deemed insignificant. Results that are slightly significant* have been highlighted.
*note: slightly significant has been interpreted as having a p-value around 0.25
In this second table control variables have been added to enhance the quality of the regression. the areas that are marked red are due to missing data, for these countries the regression has been ran with the variable which were available. Again a hump-shaped form between inequality and GDP growth takes place for every country analysed. The R2 for most regression has been improved, which indicates that a valuable variable has been added to the regression. The variable for the Composite Leading Indicator is mostly significant and positive. We can also see that the variable for social policies
B_0
B_1
B_2
B_3
B_4
B_5
R2
N
Belgium
559,11 (0,45)
(5097,11) (0,35)
9337,56 (0,34)
0,07 (0,737) (0,02) (0,63) 1,38 (0,01)
0,83
9
Canada
3518,73 (0,67)
(23139,02) (0,64) 36232,61 (0,639) 0,19 (0,40)
1,77 (0,00)
0,7 11
Czech rep.
3002,10 (0,40)
46447,02 (0,38)
46447,02 (0,377) 0,12 (0,61) (0,51) (0,19) 1,28 (0,01)
0,88
9
Estionia
(225,92) (0,82)
308,14 (0,96)
(7,01) (0,99)
0,02 (0,90) 0,02 (0,95)
1,26 (0,00)
0,95
9
Finland
5238,55 (0,49)
38533,84 (0,50)
(72647,26) (0,51) 0,51 (0,43) (0,01) (0,94) 1,31 (0,38)
0,16 12
Greece
(11123,23) (0,06) 65560,49 (0,063) (96683,55) (0,06) 0,23 (0,58) (0,62) (0,07) 0,42 (0,70)
0,87
9
Iceland
613,61 (0,53)
(4925,083) (0,51) 9059,85 (0,50)
0,72 (0,36) 1,96 (0,27)
0,01
9
Ireland
1317,97 (0,60)
(9801,56) (0,56)
16267,66 (0,55)
(0,42) (0,61) 0,00 (0,99)
1,58 (0,16)
0,49
9
Italy
(9487,68) (0,68) 1622,80 (0,82)
(2155,09) (0,85)
(0,11) (0,50) (0,05) (0,17) 1,92 (0,01)
0,88
9
Luxembourg 2990,25 (0,31)
(22299,57) (0,31) 38342,47 (0,32)
0,57 (0,32) 3,57 (0,31)
-0,04
9
Poland
(132,00) (0,44)
(72,49) (0,95)
81,33 (0,64)
0,25 (0,612) (0,03) (0,63) 1,52 (0,13)
0,3
9
Portugal
(291,41) (0,50)
1232,12 (0,63)
(1628,96) (0,64)
(0,44) (0,12) (0,07) (0,38) 0,63 (0,14)
0,87
9
Slov. Rep.
258,70 (0,70)
(2979,26) (0,53)
5437,17 (0,54)
(0,75) (0,21) (0,10) (0,39) 1,63 (0,09)
0,4
9
Slovenia
(307,55) (0,90)
2556,817 (0,89)
(6336,36) (0,87)
(0,04) (0,88) (1,10) (0,21) 1,30 (0,03)
0,81
9
has a negative effect on GDP for nine out of the fourteen countries, when only looking at the slightly significant results this is three out of four. From this we could conclude that policymakers who focus heavily on subsidies and social policies only hamper the growth of their economy. A link can be drawn as to whether this means that redistributive polices have a negative effect on the growth of an economy. It could be that because of these redistributive polices, take away money from capitalists who were planning to invest this in risky assets with relatively high returns. It should be noted
however that the variable used is not a perfect estimator for redistributive polices and that the sample size is limited.
Conclusion
The analysis of this paper has not yielded an unambiguous answer as to whether inequality is beneficial or detrimental for growth. The effect seems to be different for each country and for each level of inequality. This could be because each country is in different stage of development and thus different levels of inequality are either harmful or advantageous for this country. We can speculate with limited certainty that the relationship between inequality and GDP growth is hump-shaped, which means the effect on GDP is different for different levels of inequality. Every country in the analysis shows this non-linear relation, even though some of the results were deemed insignificant. After running the second regression these results were corroborated, also some light was shed on how subsidies and social policies affect the GDP growth. It was shown that for most countries the effect of social policies on GDP was negative. Thus it may be speculated that redistributive polices slow down the growth of an economy, a reason for this might be that policies like this disrupt investments in capital, which typically yields higher results. It must be noted however that the data and number of observations used was limited, so further research is needed to confirm these findings.
References
Barro, R. J. (2000). Inequality and Growth in a Panel of Countries. Journal of economic growth, 5(1), 5-32.
Bhattacharya, J. (1998). Credit market imperfections, income distribution, and capital
accumulation. Economic Theory, 11(1), 171-200.
Bell, L. A., & Freeman, R. B. (2001). The incentive for working hard: explaining hours worked differences in the US and Germany. Labour Economics, 8(2), 181-202.
Cingano, F. (2014), “Trends in Income Inequality and its Impact on Economic Growth”, OECD Social,
Employment and Migration Working Papers, No. 163, OECD Publishing.
http://dx.doi.org/10.1787/5jxrjncwxv6j-en OECD
Cornia, G. A. (2004). Inequality, growth, and poverty in an era of liberalization and globalization (No. 4). Oxford University Press on Demand.
Deininger, K., & Squire, L. (1996). A new data set measuring income inequality.The World Bank
Economic Review, 10(3), 565-591.
Deininger, K., & Squire, L. (1998). New ways of looking at old issues: inequality and growth. Journal of
development economics, 57(2), 259-287.
Galor, O., & Zeira, J. (1993). Income distribution and macroeconomics. The review of economic
studies, 60(1), 35-52.
Korpi, W., & Palme, J. (1998). The paradox of redistribution and strategies of equality: Welfare state institutions, inequality, and poverty in the Western countries. American sociological review, 661-687.
Kuznets, S. (1955). Economic growth and income inequality. The American economic review, 45(1), 1-28.
Lin, Y. C., Huang, H. C. R., & Yeh, C. C. (2014). Inequality-growth nexus along the development process. Studies in Nonlinear Dynamics & Econometrics,18(3), 237-252.
Iyigun, M. F., & Owen, A. L. (2004). Income inequality, financial development, and macroeconomic fluctuations. The Economic Journal, 114(495), 352-376.
Tabassum, A., & Majeed, M. T. (2008). Economic growth and income inequality relationship: role of credit market imperfection. The Pakistan Development Review, 727-743.
Thornton, J. (2001). The Kuznets inverted-U hypothesis: panel data evidence from 96 countries. Applied Economics Letters, 8(1), 15-16.
Voitchovsky, S. (2005). Does the profile of income inequality matter for economic growth?. Journal of
Economic growth, 10(3), 273-296.
World Bank. 2015. World Development Indicators 2015. Washington, DC: World Bank. doi:10.1596/978–1-4648–0440–3. License: Creative Commons Attribution CC BY 3.0 IGO
